The E-bus alt feed: a different idea

[Southern Pete]

The spilt bus idea is designed to cope with an airborne alternator failure in a single alternator, single battery aircraft. Soon after an alternator failure is identified 2 switch selections will vastly reduce the electrical load and give the pilot plenty of time to figure out what to do next. Eliminating the alternator field and master relay current can save between 3 and 5 amps, on top of the high power services, lights, heated pitot, etc.

Flying with the direct feed from the battery to the endurance bus on will mean it will be left on at some point, draining the battery. Normal use is master only on. The e-bus feed can be used as a "Clearance Delivery" switch, depends how it is wired up.

There are plenty of other methods to implement this scheme, including switches instead of the diode. It will achieve the same end, but will need some careful labeling to make sure operation is clear when the alternator fails 5 years after the aircraft was built.

There are draw backs to using a diode, mainly the voltage available to the e-bus side of the bus is around a volt lower than the main bus. Usually that is no factor, but it can be significant. Reliability usually comes with low parts count which is why the diode scheme was invented. With a second alternator the need to quickly shed the high power loads is less important so the e-bus layout is less useful. For brown-out protection something like the TCW Intelligent Power Stabilizer can be useful.
Pete

[rapid_ascent]

I personally I think the addition of a capacitor is just adding unneeded components. The most common failure of capacitors is shorting so it will effectively take out the bus it?s connected to if it fails.

My planned sequence is slightly different from Joes but it?s similar in that there is only one switch operation required in case of an issue.

* turn on ebus to get ATIS
* turn off ebus to protect avionics during start
* turn on main alt and do engine start
* turn on ebus and I?m in my normal run state

During this last step I don?t expect any voltage spiking. The load is being supplied by the main bus. It will be a diode voltage drop lower than the battery voltage. Let say between 0.7 an 1.0 volts. Then when the ebus is switched on the ebus voltage will rise to the battery voltage.

In the case of an alternator failure I switch off the main bus. The voltage will stay constant since the diode in normal operation is reversed biased (turned off). This is due to the ebus voltage being higher than the main bus voltage by the diode drop.

[avrojockey]

Originally Posted by N546RV

So I've been thinking about the Nuckolls E-bus setup, specifically the handling of having an alternate feed off an always-hot battery bus in case of a failure affecting the main bus. Bob's setup takes the same basic form everywhere I've seen it: the main bus feeds the E-bus through a diode, and the alternate feed is provided through a switch and/or relay (depending on expected E-bus loads).

The reasoning behind the diode between the main and E-bus is to prevent the E-bus from feeding the main bus and possibly popping a fuse/breaker. Bob's assertion seems to be based on an alternative scenario where two switches would be provided, and it would be incumbent on the pilot to ensure the main->E-bus feed switch was opened before closing the E-bus alt-feed switch, to prevent the aforementioned backfeeding.

However, it seems this issue can be fairly easily solved with careful switch choices, such that it's not possible to have both feed paths closed simultaneously. For example, if a five-pin relay is used for feed switching, the main bus could feed through the normally-closed contact, and the batt bus through the normally-open one. The choice of how to feed the bus is still controlled by a single switch throw, the only difference being that that switch position definitively disconnects one feed while connecting another one.

Example diagram:



This seems elegantly simple enough that I'm immediately suspicious that I'm missing some downsides. Some thoughts that come to mind in this vein:

• There's the possibility of the relay somehow failing such that both contacts close, thus possibly blowing the batt bus fuse and killing the E-bus in an emergency. Not sure what the likelihood of this is.
• If the relay somehow fails with both contacts open, the E-bus is now dead. This question seems to come down to the relative chance of failure between the diode and the relay.
• Switching feed paths means momentarily cutting power to the E-bus. This seems inconsequential since, when switching to the E-bus, presumably power has been lost already.

Any other detractions I'm missing? Any thoughts on the concerns above?

I have a similar setup with a slight variation...I have a guarded DPDT On-On switch that controls the feed for the E-BUS it's either from the main bus or straight off the battery via relay and fuse link. One pole of the switch controls the source of power and the other controls the relay. A single throw of the switch moves power source from the main bus to the hot battery.

The fuse link off the battery is 15A, normal power to off the main bus is protected by 15A fuse, and I limited the EBUS load to 10A.

This design was not for endurance as much as it was for emergencies when attempting to isolate an electrical fault through bus isolation...I have a all electrical IFR cockpit with a single alternator and a single battery and the fuses are not accessible in flight.

[petercavallo]

I planned for 1 failure back up, Alt goes bad , ebus on with 1 switch (no relays no big thinking, see the blinking alt light and do something)
Same with having one LSE ignition and 1 mag one bad use the other
Mech fuel pump bad turn electric on
No regarding the engine .......Glide

So far so good, but always worth looking at new ways to do things

Peter

[walkman]

Originally Posted by Southern Pete

Flying with the direct feed from the battery to the endurance bus on will mean it will be left on at some point, draining the battery. Normal use is master only on. The e-bus feed can be used as a "Clearance Delivery" switch, depends how it is wired up.

Hasn't been a problem in almost 8 years.

Perhaps I'm more careful.

Its part of my shutdown procedure, and I always eyeball the switches after exiting the aircraft.

Its easy to tell anyway, as my GTN650 would be on. That's a pretty big annunciator.

[walkman]

Originally Posted by rapid_ascent

I personally I think the addition of a capacitor is just adding unneeded components. The most common failure of capacitors is shorting so it will effectively take out the bus it?s connected to if it fails.

My planned sequence is slightly different from Joes but it?s similar in that there is only one switch operation required in case of an issue.

* turn on ebus to get ATIS
* turn off ebus to protect avionics during start
* turn on main alt and do engine start
* turn on ebus and I?m in my normal run state

During this last step I don?t expect any voltage spiking. The load is being supplied by the main bus. It will be a diode voltage drop lower than the battery voltage. Let say between 0.7 an 1.0 volts. Then when the ebus is switched on the ebus voltage will rise to the battery voltage.

In the case of an alternator failure I switch off the main bus. The voltage will stay constant since the diode in normal operation is reversed biased (turned off). This is due to the ebus voltage being higher than the main bus voltage by the diode drop.

I turn on ALT/MASTER, EBUSS, E-IGN, MAG, Fuel Pump (as necessary). The switches are in that order furthest to nearest.

My MAG is OFF-ON-MOM so to start I simply push the MAG switch to the right into its MOM position. That forces me to have the MAG on to start and pretty difficult to turn it off again by accident although I'd catch it on runup.

Immediately check OP and RPM on the EIS.

The next two switches are EFIS #1 and then #2 so they go on once the engine is running, then I pick up ATIS and do other checks while the EFIS's are coming online.

My switches are arranged in groups and ordered such that in most operation I'll turn them on/off in the order they are mounted. Works well for me, simple and pretty failsafe.

[RV6_flyer]

Originally Posted by N546RV

So I've been thinking about the Nuckolls E-bus setup, specifically the handling of having an alternate feed off an always-hot battery bus in case of a failure affecting the main bus. Bob's setup takes the same basic form everywhere I've seen it: the main bus feeds the E-bus through a diode, and the alternate feed is provided through a switch and/or relay (depending on expected E-bus loads).

The reasoning behind the diode between the main and E-bus is to prevent the E-bus from feeding the main bus and possibly popping a fuse/breaker. Bob's assertion seems to be based on an alternative scenario where two switches would be provided, and it would be incumbent on the pilot to ensure the main->E-bus feed switch was opened before closing the E-bus alt-feed switch, to prevent the aforementioned backfeeding.

However, it seems this issue can be fairly easily solved with careful switch choices, such that it's not possible to have both feed paths closed simultaneously. For example, if a five-pin relay is used for feed switching, the main bus could feed through the normally-closed contact, and the batt bus through the normally-open one. The choice of how to feed the bus is still controlled by a single switch throw, the only difference being that that switch position definitively disconnects one feed while connecting another one.

Example diagram:



This seems elegantly simple enough that I'm immediately suspicious that I'm missing some downsides. Some thoughts that come to mind in this vein:

• There's the possibility of the relay somehow failing such that both contacts close, thus possibly blowing the batt bus fuse and killing the E-bus in an emergency. Not sure what the likelihood of this is.
• If the relay somehow fails with both contacts open, the E-bus is now dead. This question seems to come down to the relative chance of failure between the diode and the relay.
• Switching feed paths means momentarily cutting power to the E-bus. This seems inconsequential since, when switching to the E-bus, presumably power has been lost already.

Any other detractions I'm missing? Any thoughts on the concerns above?

Downside:
Contacts will pit and fail. Failure mode could be that it does not switch.

For 20-amps or less, I would rather use the diode and a switch instead of the relay.

In my career before retirement, I have seen more relays fail than I have switches.

[supik]

That's a clever flow!

Originally Posted by walkman

I turn on ALT/MASTER, EBUSS, E-IGN, MAG, Fuel Pump (as necessary). The switches are in that order furthest to nearest.

My MAG is OFF-ON-MOM so to start I simply push the MAG switch to the right into its MOM position. That forces me to have the MAG on to start and pretty difficult to turn it off again by accident although I'd catch it on runup.

Immediately check OP and RPM on the EIS.

The next two switches are EFIS #1 and then #2 so they go on once the engine is running, then I pick up ATIS and do other checks while the EFIS's are coming online.

My switches are arranged in groups and ordered such that in most operation I'll turn them on/off in the order they are mounted. Works well for me, simple and pretty failsafe.

[N546RV]

Thanks for all the commentary, folks. After reading all this, I've decided to stick with the diode; overall, I think the talk about failure rates is what really put me over the line.

I'm also liking the idea of having the alt feed switch normally on; one of the things that made me hesitant to use the diode was the voltage drop. I think Bob Nuckolls's argument that the drop is immaterial holds water in a traditional single-alternator system, but I'm planning on a primary and standby alternator, with the standby regulator set lower so it comes online automatically in the event of a failure of the primary.

So in a case where I'm running on the standby, now my E-bus ends up down around 12.5 volts - everything will continue to run, but one critical point is that I'm close to the threshold where the Skyview system won't charge/maintain its backup battery. Probably not a huge deal real-world, but enough to make me consider options.

Anyway, the tactic of having the alt feed normally on makes the diode voltage drop pretty immaterial, outside of a scenario where the master contactor is open. I also like that it makes yet another in-flight incident response require no direct action from me.

I'm getting pretty close to being satisfied with my power distribution architecture.

(Note: I'm building an electrically-dependent airplane, hence my somewhat elevated interest in designing a fault-tolerant system)

[rapid_ascent]

Philip,

Glad we were able to convince you. In general simpler is better.

Having said that... There is a circuit referred to as a virtual diode. It uses one or more MOSFETs and a control circuit nstead of an actual diode. This results in very little voltage drop. The voltage drop is dependent on the current and how good of a MOSFET you use. These days you can buy MOSFETs which can handle a couple hundred or more amps and have very low in resistance.

I built a circuit card with one of these Virtual Diode circuits for my EBus feed. One of the other advantages of this circuit is the MOSFET can be switched off and on so that it can also perform the function of a solid state relay. Anyway when I can get some time or when I actually need it I?ll finish validating my circuit.