What's new
Van's Air Force

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

Alternator Voltage With Battery Disconnected

DanH

Legacy Member
Mentor
Theory says an alternator should supply electrical power with the battery disconnected (i.e. master switch "OFF" or broken coil ground, master contactor open). It's a fundamental for electrical system design to support the engine controls and avionics we're installing. Some folks even go dual alternator and one battery.

I know it should work that way, but what really happens?

Sunday evening, a quick test, Plane Power 60A, 4500 feet over 08A. Open the avionics master (yes, I have one) to disconnect the radio, transponder, GPS, intercom, and autopilot...just a caution. Open the master switch. EFIS remains up, as it is connected to the main bus, which remained connected to the alternator. Both electronic ignitions remain up, as they have their own battery-direct feeds.

The observation? The digital voltage readout on the EFIS began cycling rapidly, almost as fast as the eye could see, in a range from roughly 13 to 16 volts. I say "roughly" because I have literally have no precise way to know how high it was spiking, or how low. It simply cycled too fast for the display to keep up.

With the avionics offline, only the GRT Sport was exposed to the cycling voltage. Didn't miss a beat, but I did not open its aux battery input, so I don't know if the primary power input voltage cycled low enough to cause a reboot.

It would also be nice to know how high the voltage is actually going. Anyone have a laptop with the necessary input to record a similar test at a high sample rate? I guess I could buy something, but nothing I already have handy will do it. The extension cord for my old o-scope is definitely not long enough ;)
 
Last edited:
Normal?

I must disagree, I think this cycling may be normal without a battery. With no battery, and the avionics bus disconnected, the alternator's control loop is all messed up ( a technical term). Just my opinion, but I don't think it is wise to run an alternator without some kind of battery or large capacitor, absorbing the fluctuations. Normally I think all the electronic boxes may have a combined capacitance high enough for the system to be OK with the battery off line, but with the master off, then these are also disconnected.
 
The cycling is normal. An alternator counts on the battery to smooth out the voltage fluctuations.
Usually the squeal is so loud when the master solenoid opens that you can't hear anything. Good thing you left them off. :)

Vic
 
Another worthwhile experiment Dan. Hope you continue on this some more. Wouldn't this be easier/safer on the ground with a test rig of sorts?

I'm certainly interested to see the results.
 
I'll check my old Fluke.
Mine has two settings for reading peaks. I don't have the manual handy but I think the normal bandwidth for MIN/MAX is 1 mS. There's a mode I can select which will acquire 250 uS peaks. Your mileage may vary.
 
There is possibly something else we can learn from this. If an owner (or a phase 1 flyer) sees fast varying voltage, the connections leading to the battery, including master solenoid, may be root cause vs charging system commonly eyeballed.
 
Passing up the opportunity to buy a new tool? Not Dan!!!

You can get a little cigarette-pack sized O-Scope these days that plugs in to a laptop for its display. Not that expensive. I haven't bought one because I haven't had a need - but used that pin the old job. Sounds like now you have an excuse!

Paul
 
I conducted an experiment similar to Dan's. After leveling off in cruise, I shut off the battery contactor, but left everything else turned on including avionics and alternator. Technically speaking, the permanent magnet alternator in the Rotax 912ULS is called a dynamo. It is rated at about 20 amps maximum. Without the battery, system voltage varied between 13.25 and 14.7 volts or 3/4 volt plus and minus from normal. All electrical loads continued to operate normally. But the intermittent voltage peaks set off the high voltage alarm on the Dynon D-180. After landing, I downloaded the D-180 data to analyze using MS Excel. I agree with others that the battery is needed to stabilize the electrical system voltage. I had posted this experiment on the Matronics AeroElectric List in July of 2016.
 
Maybe this is why the "Cessna Style" Master/Alternator switch doesn't allow turning off the battery without also turning off the alternator....... :confused:

Other than testing or curiosity, what set of circumstances would prompt one to turn the battery off and leave the alternator online?
 
Hmmm.

My thought is that conclusions, like sharks, are best not jumped.

I'd be doing a little logic flow chart. If the alternator has OV protection built in (as claimed on the web site), then either the voltage never went to an unsafe level, or the alternator's OV protection doesn't work. (As an aside, I'd love to see how the built-in overvoltage protection is implemented.)


If the alternator's regulator senses output voltage at the B lead terminal (based on the published wiring diagram, it seems likely), there should be no difference in regulation with or without the battery. But there *will* be some ripple in the alternator's output, and the battery would act as a smoothing element (like a capacitor)

And there's the question of what's meant by the installation cartoon at
http://planepower.aero/pdf/AL12_EI60%20Installation.pdf
What's the actual wiring of the enable switch? One wire from the 5A fuse to the red switch, and two wires leaving, tied to two out of three terminals in that connector. What do those two terminals do?

Pure speculation, but if the field winding is brought out to that switch, could you be getting an oscillation in the regulator circuit? Precedent would be some older certified a/c using external regulators. If you've ever seen the voltmeter oscillating by a couple of volts during operation, that's been attributed (by Bob Nuckolls, IIRC) to resistance in the regulator's sense & field wiring, causing oscillations.

Charlie
 
Maybe this is why the "Cessna Style" Master/Alternator switch doesn't allow turning off the battery without also turning off the alternator....... :confused:

Other than testing or curiosity, what set of circumstances would prompt one to turn the battery off and leave the alternator online?

With a battery like an EarthX, which has its own ability to disconnect, you might not have a choice.
 
Since the crowbar protection circuit shorts the field to ground on over voltage. I would be curious to know what the amp draw of the field circuit is during the testing. Could the voltage hunting be caused by the crowbar protection dropping in and out, or just the lag in response in the regulation circuit with no battery to act as a damper.:confused:

Best regards,
 
I know from previous automotive experience that running a car on alternator only without battery will fry the alternator. Embarassing story, but killed 2 new alternators this way. According to the test equipment, the voltage regulator basically failed full open, with alternator putting out very high voltage.

Caleb
 
Yeah, and thinking on it a direct short to a 5 amp cb wouldn't take to long to trip...:rolleyes:

Regards,

Actually, with no external source of power, the alternator is the only thing supplying power to its own field, so shorting the field via crowbar should instantly shut down the alternator. I say 'should' since there may be enough residual magnetism to restart the alternator.

Another way to add a buffer to the battery-less electrical system is to turn on a landing light... but that would probably collapse the field enough momentarily to shut down the alternator too.

Either way, would make a nice science experiment.

Heinrich
 
A pretty good picture of what you can expect to see with an O-scope:

I personally added a couple of large capacitors to my main and avionics busses to help smooth things out in case this scenario ever occurs (battery disconnected from bus due to master relay failure).

123GT_1.jpg
 
Walt's images may well help explain what the EFIS' voltmeter was saying.

Most digital voltmeters (unless you pay for an RMS feature) simply sample the voltage on their input, at some interval (typically several times a second), and update the display with each sample. With the alternator's output 'unfiltered', some samples will be at the peak, some will be somewhere on the ramps, and some samples will be in the valleys.

Without the battery, there may be somewhat more variation in DC level, since the regulator is almost certainly chasing that ripple, which wouldn't be there with a battery or capacitor to smooth it. But it seems unlikely that it ever went more positive than what's safe, or the OV protection circuit should have fired.
 
I know from previous automotive experience that running a car on alternator only without battery will fry the alternator. Embarassing story, but killed 2 new alternators this way. According to the test equipment, the voltage regulator basically failed full open, with alternator putting out very high voltage.

Ahhh, precisely why I got curious. In the past we built sport airplanes which remained flying with no electrical system at all. Electrical redundency was for IFR, something to keep the avionics running. Now we're moving towards EFI and EI. Perhaps we should check some fundamental assumptions...like being sure about what really happens after a battery disconnect, for example.

A pretty good picture of what you can expect to see with an O-scope

....when connected to a "normal" system, i.e. battery connected....a relatively stable output voltage (bottom two illustrations). With the battery disconnected, it's oscillating between unknown voltage values. Or at least mine is oscillating. That's why I'd like to see someone measure another Plane Power, as well as a B&C charging system.

I personally added a couple of large capacitors to my main and avionics busses to help smooth things out in case this scenario ever occurs (battery disconnected from bus due to master relay failure).

Thank you Walt. Now to the meat of the matter. Have you ever measured the bus voltage with and without the capacitors, battery disconnected?
 
Thank you Walt. Now to the meat of the matter. Have you ever measured the bus voltage with and without the capacitors, battery disconnected?

Nope, I know that a voltmeter wouldn't really tell me much and I don't own a portable O-scope.
 
Data acquisition systems like these work well; I've used two for various home/work purposes. I have the DI-149, which has been replaced by the DI-1110. (I bought it several years ago to track down why my car was destroying batteries every year.) The DI-1100 for $59 will do what you need, I believe. USB to a laptop (real time), and it's easy to export to excel later if you wish. Software not the friendliest, but easy once learned.

I'm sure there are others out there as well.
 
The DI-1100 for $59 will do what you need, I believe. USB to a laptop (real time), and it's easy to export to excel later if you wish. Software not the friendliest, but easy once learned.

Hey, you been talking to Paul? It's not fair to tempt a tool junkie ;)

For penance, you are hearby required to go forth and measure your bus voltage, master OFF.
 
Hey, you been talking to Paul? It's not fair to tempt a tool junkie ;)

For penance, you are hearby required to go forth and measure your bus voltage, master OFF.

The answer will be zero volts with the master off, since my alternator will necessarily go off when I turn master off (alt is full up position of master)...
 
Why not just measure AC content of the buss?

While it is hard for the normal DVOM to measure DC ripple, it can measure AC - There is an AC content, so why not just measure with the DVOM on AC volts and do that with the battery in and out of the circuit?

It will not get the peaks, but you will get some comparison. This will also make it easier for others to provide data.
 
Bill is right. Whenever I suspect a bad alternator I put the DVM on AC and measure "ripple". Quick way to find a bad diode in an alternator.

Vic
 
With a battery like an EarthX, which has its own ability to disconnect, you might not have a choice.

The EarthX BMS does not disconnect for an over voltage situation. The only time the battery will disconnect is when it has been completely drained to about 2% energy left in it from an over discharge situation. I am copying directly from the manual how this safety feature works:
In the event of a charging system failure where the voltage increases to above 15.5V, the resistance to charging current increases, and above 16V the charging current is completely blocked. The time delay for this feature is 1 second to allow the aircraft alternator?s over voltage protection (crowbar circuit) to activate first. This design offers charge voltage protection greater than 40V. The discharge current (current out of battery) is unaffected in this situation. EarthX strongly recommends having automatic over-voltage protection (crowbar) for alternator type charging systems (not required for <20 Amp pad mount standby alternators).
 
The EarthX BMS does not disconnect for an over voltage situation. The only time the battery will disconnect is when it has been completely drained to about 2% energy left in it from an over discharge situation. I am copying directly from the manual how this safety feature works:
In the event of a charging system failure where the voltage increases to above 15.5V, the resistance to charging current increases, and above 16V the charging current is completely blocked. The time delay for this feature is 1 second to allow the aircraft alternator?s over voltage protection (crowbar circuit) to activate first. This design offers charge voltage protection greater than 40V. The discharge current (current out of battery) is unaffected in this situation. EarthX strongly recommends having automatic over-voltage protection (crowbar) for alternator type charging systems (not required for <20 Amp pad mount standby alternators).

You have said this before and, frankly, it does not make sense to me. How can the battery be CONNECTED and CURRENT BLOCKED at the same time. A little more technical explanation is in order here please.
 
My take

You have said this before and, frankly, it does not make sense to me. How can the battery be CONNECTED and CURRENT BLOCKED at the same time. A little more technical explanation is in order here please.

Does the Earthx go off line if the voltage gets too high to protect itself, and then will restore power to the aircraft once the voltage comes within normal range again?

If this is true, then the earthx fundamentally chnages to aircraft's electrical system. With the traditional battery, the aircraft electrical system can be thought of a " bus on battery" design so equipment is always at battery voltage, good or bad.
With the Earthx deciding itself to go off line or on line, then this is no longer the case and it is unclear what is the master control of the bus voltage. Could it that this is now an "Bus on alternator"design? If so probably needs two alternators for IFR redundancy, and caps to control alternator ripple if batery goes off line?

Update: seems like Earth X has taken care of this, see post below.
 
Last edited:
The simple fact that the EarthX has a baked-in feature that *can* take it off-line means that that possibility must be considered in the a/c's power architecture. How reliable that 'feature' is, should not be part of the calculus.

Just because the battery's protection circuit isn't *supposed* to take it off line except at discharge, doesn't mean it *can't* take it off line for other reasons, like a malfunction in that protection circuit.

So... determining whether dropping the battery out of the circuit matters to the rest of the system would be the 1st step.

It may be that it just causes a 1 volt ripple in the DC from the alternator. That might (or might not) be acceptable.

It may be that a fat capacitor would solve any ripple issues. But then any maintenance personnel would need to know about its existence, to avoid injury or damage to other components during maintenance.

Etc etc.
 
If I remember correctly we recently had an occurance of an out of control voltage excursion reported here on VAF which took out the whole panel due to a out of control alternator with no OV protection and EarthX disconnect.
 
It may be that a fat capacitor would solve any ripple issues. But then any maintenance personnel would need to know about its existence, to avoid injury or damage to other components during maintenance.

Small low voltage electrolytic caps will leak down pretty quick on their own and basically store so little energy they are virtually harmless. Don't think I've every heard of anyone being killed or injured by 12v.
 
Small low voltage electrolytic caps will leak down pretty quick on their own and basically store so little energy they are virtually harmless. Don't think I've every heard of anyone being killed or injured by 12v.

Depends on your definition of 'small', and where you listen.

Edit: Above typed in a hurry before running some errands; I should have been more specific. The problem with 'small' is determining how small you can go & still get effective filtering for the job. For example, the filters commonly used with 'dynamo' style PM alternators are big enough to at very least, cause a major 'startle' reaction if an unsuspecting person shorts one (or a wire connected to it) to ground. And good electrolytics will store a charge much longer than you might expect, unless there's something in the circuit to 'bleed off' the charge.

Charlie
(no Holiday Inn last night, but a couple of careers in electronics)
 
Last edited:
If I remember correctly we recently had an occurance of an out of control voltage excursion reported here on VAF which took out the whole panel due to a out of control alternator with no OV protection and EarthX disconnect.

To avoid muddying the waters, it should be pointed out that even a conventional chemistry lead-acid battery will only slow the climb of an OV event by a some small number of seconds; not stop it. If it's not detected soon enough or mitigated by an OV protection circuit, it will likely damage avionics *and* the battery.
 
You have said this before and, frankly, it does not make sense to me. How can the battery be CONNECTED and CURRENT BLOCKED at the same time. A little more technical explanation is in order here please.

Hi Bill,
The EarthX 100 series batteries were designed specifically for aircraft, with aircraft electrical systems and safety concerns in mind.

Yes, an EarthX ETX680, 900 and 1200 can block charge current independent of discharge current. It can also block discharge current independent of charge current. This is done with two Mosfets in a back to back configuration. One Mosfet is for charge control and one is for discharge control. Furthermore, the there are two independent charge and discharge current paths; redundant current paths. And for additional protection, the discharge Mosfets default ?ON?. That is, in the event of a micro-processor failure the discharge Mosfets stay on. The EarthX 100 series batteries meets and exceed the requirements for a certified aircraft battery as defined by FAA?s DO-347, DO-160 and DO-254.
 
Good design

Seems like the earthx doesnt rely on software to allow discharge to take place. So if the altenator goes off line for whatever reason, and the earthx software screws up, the earthx will still power the aircraft. This is a good design and needs to be included in the electrical system architecture.
 
Doesn't matter what kind of battery. They can all get disconnected...terminal corrosion, or any open in the master contactor's coil circuit.

Question here is "What really happens when it gets disconnected?"
 
Last edited:
Be Careful.....

Be careful disconnecting the battery from the electrical system while the alternator is running. If the battery is taking a charge and you take it out of the loop, you will in effect create a load dump, potentially sending 100+ volts through your electrical system for up to 400 ms.

I didn't know this could happen until it happened to me. $10K later and a new panel, I am a wiser man.

https://en.wikipedia.org/wiki/Load_dump
 
Interesting. One of the tests we've been performing for years on our ECUs is to pull the positive battery cable off while the engine is running. We've done this dozens of times over the years and nothing has ever fried. ND internal reg alternator.
 
Back
Top