Roee Kalinsky's RV-7A Project

Main DC Power Distribution
Home Up News Specifications

 

Back Up Next

Main DC Power Distribution

Running Total Hours: 0.0

 

2009.05.27: (0.0) The early RV-7A used a Concord RG-25XC battery mounted on the left side of the firewall (DWG 31).  This is a 12 V, 24 Ah absorbed glass mat (AGM) recombinant gas valve regulated lead acid (VRLA) aircraft battery, and weighs 23.5 lb.  The RV-7 on the other hand used an Odyssey PC680 battery mounted on the right side of the firewall (DWG 31A), I'm guessing because the Concord battery installation would interfere with the main landing gear mounts in the tail dragger configuration.  This is a 12 V, 17 Ah AGM VRLA battery, and weighs 15.4 lb (that's 8 lb lighter, and also considerably more compact than the Concord).  Odyssey does not explicitly market it as an aircraft battery (I believe it's commonly used in motorcycles, snowmobiles, jetskis, etc.), but it has characteristics that are very well suited for aircraft use, and it has a long and excellent service record in many RV's and other experimental aircraft.  Originally, Van's made the Odyssey battery available as an optional replacement for the Concord on the RV-7A, but they've now become standard on both aircraft.  I would have chosen the Odyssey battery in any case for my airplane.

The battery box hardware normally comes as part of Van's firewall forward kit, but I just bought it separately now since this was a convenient time for me to install it (I'll delete it from my firewall forward kit when the time comes).  So I'm working on the battery box.

2009.12.19: (0.0) [Some of this entry is quite old] Finished fitting the battery box, installing its three nutplates, and cutting the eight lightening holes shown in the plans.  Cutting the lightening holes took a long time and completely dulled down two hole saws.  A punch would likely have worked much better, but I didn't have one.  Anyhow, I cleaned up the lightening hole edges, removed a bit of rust from the inside surfaces of the box, and gave the entire box a coat of AKZO epoxy primer.  The battery box came powder coated from Van's, but: 1. The lightening hole cuts need to be coated for corrosion protection, and 2. As shipped from Van's, the inside surfaces of the battery box didn't have good powder coat coverage and were already starting to rust.  Now with a coat of AKZO it should last a good long time.

2011.03.07: (0.0) Fitted the various main DC power distribution components that will mount on the forward side of the firewall.

Main battery

The main battery is an Odyssey PC680, and the battery box has been fitted at the standard location as described above.

Master and starter contactors

The master and starter contactors (Cole Hersee p/n 24115 and 24021 respectively, Van's p/n ES 24115 and ES 24021) have been fitted at the standard locations.  Although, I made the firewall doubler a little bigger than Van's drawings to span all the way between the two angle stiffeners (trying to minimize firewall drumming).

Ground power connector and contactor

Fitted the ground power connector and contactor, which is my own customization, not found in Van's stock design.  See details here.

Main alternator and main buss fuses

Breaking from Van's tradition here, going in a different direction inspired by the AeroElectric Connection.

Tradition is to run the main feed (master contactor output) and the alternator B-lead into the cockpit to meet at the breaker panel.  The two big disadvantages of this approach are 1) bringing all that alternator noise into the cockpit, and 2) risk of unstoppable in-cockpit electrical fire in case of hard short in the alternator B-lead or main feed wire forward of the breaker panel.

The alternate approach I'm taking in order to eliminate these problems is to keep the alternator B-lead entirely on the forward side of the firewall, and provide circuit protection against hard shorts for the alternator B-lead as well as for the main feed to the cockpit via ANL fuses on the forward side of the firewall.  The obvious disadvantage of this alternate approach is that if a fuse does blow, it is not resettable/replaceable in-flight.  But the characteristics of ANL fuses are such that a "nuisance trip" is virtually impossible (unlike typical alternator B-lead circuit breakers).  And furthermore, in the event that a hard short does occur, there is little or no value to being able to reset or replace the circuit protection device in-flight.  The advisable course of action would be to make use of the aux electrical system to safely get on the ground, and only then troubleshoot the fault in the comfort and safety of a hangar.

So... I mounted the two ANL fuse blocks (Bussman p/n 4164, sold by B&C as p/n C903-1) below the master contactor, to minimize the length of the unprotected main feed wire between the master contactor and the ANL fuses.  Note that these fuse blocks each mount with a pair of 3/16" countersunk screws (not provided), and the countersink angle in the fuseblock is 83 degrees, which is standard in commercial hardware, not the 100 deg countersink angle that's standard in aircraft hardware.  I used 1 inch spacing (center-to-center pitch) betweent he two fuse blocks.

The fuses themselves are 60A rated (Bussman p/n ANL-60, sold by B&C as p/n C905), but it's important to note the actual characteristics, shown in the Bussman datasheet as a time-current curve.  The ANL-60 will actually sustain upwards of 100A indefinitely, somewhere around 200A for one second, about 600A for a tenth of a second, and are ultimately rated to safely interrupt up to 2,700A, which is more than even a dead short across the battery can produce.  Bottom line is that these fuses will only ever blow under a hard short, which is the one and only time we want them to blow.

Note also that these are sometimes referred to as "current limiters", which to me (coming from the electronics world) means something very different.  As far as I'm concerned, calling an ANL a "current limiter" is misleading.  It's a fuse.  Simple as that.

Main alternator B-lead shunt

Main alternator current will be measured using an 100A, 100mV shunt.  Mine comes from Advanced Flight Systems (p/n 44105) as part of the engine sensor package, and is actually a DELTEC MKA-100-100.  But note that this is actually a standard milspec part, and is available from a variety of manufacturers and distributors.  These shunts go back many decades, and are known under several generations of specs and designations, including (in chronological order, I think...) AN3900, MS91586, type designation MSA800, and finally AA55524/1.

Firewall ground bolt

I'm deviating significantly from Van's grounding scheme.  Van's uses airframe grounding for airframe loads throughout, which is not entirely unreasonable for a metal airplane (although it does have some disadvantages, and my plan is to run dedicated ground wires).  But furthermore, Van's even uses airframe structure to conduct starter current, by virtue of tying the battery ground wire to a bolt on the firewall and tying the engine case ground wire to a different bolt elsewhere on the firewall.  I don't like that at all, for several reason related to grounding reliability in general, starter performance, and just the idea of running 200 amperes through thin aluminum and stainless steel structure and how it may heat certain parts of that structure or promote long term corrosion.

My plan is to run the three ground wires that carry starer current (engine case, battery, and external power connector) to a single bolt on the firewall.  This will provide a lower resistance path for starter current, and avoid running it through airframe structure entirely.  I will also run a fourth wire from that bolt to a ground disribution point in the cockpit, from which ground wires will fan out to the various airframe loads along with their (positive) power wires.  So the airframe is still grounded via the firewall bolt, but is not generally relied upon nor used as a return path for any significant loads (a few sensors and low-power loads still do get ground via their case mounting, unfortunately).

Note that the AeroElectric Connection also promotes a central ground point architecture using a "forest of tabs" component sold by B&C, which is a brass plate with a bunch of fast-on tabs soldered to it, and a 5/16 brass bolt and hardware for the heavy ground wires.  Not a bad idea, although I bought one of these parts and was not terribly impressed with its quality, and had some misgivings about structurally how it mounts.  And anyway, I'm planning on a PCB main box that will distribute most of the accessory grounds along with their power, so the "forest of tabs" aspect is redundant.  So bottom line, I agree with the central ground point philosophy they advocate, just wasn't crazy about how these guys implemented it, so I'm rolling my own.

Note also that unlike AeroElectric/B&C, I'm not using any brass hardware for my firewall ground point.  Brass is often used in this type of application as a compromise of physical properties, primarily between conductivity (not nearly as good as copper, not nearly as bad as steel) and structural strength (not nearly as good as steel, not nearly as bad as copper).  I'm using a standard AN4 bolt (cad-plated steel), which is very strong, and which in my grounding architecture is not called upon to carry current and therefore is not required to be a good conductor.

My firewall ground bolt (and nutplate on the aft side of the firewall) is located toward the top of the diagonal stiffener, a point that is both structurally solid and conveniently located for running the ground wires that converge there.

Welding cable

Starter current can be as high as 185A (SkyTec 149-NL spec).  This of course is only for a few seconds at a time, so some overload of wire specs by design is acceptable.  Long story short, 2 or 4 AWG wire is generally used for all wires that carry starter current.  As for wires carrying alternator current, 60A continuous puts us at 6 AWG.

Now, throughout the airframe I use almost exclusively milspec tefzel-insulated wire for its proven physical properties, especially as they pertain to abrasion and fire.  But one undesirable property of this milspec wire is that it is very stiff.  Not a problem for smaller wires.  But in the heavy gauges of the primary power distribution wires as discussed above, the stiffness becomes a real issue, as the wire can exert significant forces on the components that it attaches to.  And the problem is even further exacerbated where one end of the wire is subject to engine vibration, as is the case with the starter, alternator B-lead, and engine case ground wire.  Terminals breaking over time is a common problem in the field.  And a proven remedy is to replace the stiff milspec wire with welding cable, which is specifically designed to be very flexible while still environmentally rugged.  I'm going this way from the start.

Now unlike standardized milspec wire, welding cable is a class of commercial products that vary in physical properties from one product to the next.  So I did a little research to find the most suitable variety.  The type I settled on is called "Super Vu-Tron Welding Cable", made by General Cable under the Carol Brand.  Compared to more typical welding cable (such as General Cable's "Carolprene Welding Cable"), Super Vu-Tron is even more flexible, uses finer strands (34 AWG rather than 30 AWG), has a higher continuous current rating for a given gauge, and uses an insulation material that is more resistant to oils, solvents, and ozone.

I'm using Super Vu-Tron 4 AWG for wires carrying starter current, and 6 AWG for alternator current.  [Note that I also bought a short length of 2 AWG just to check it out.  Still quite flexible, and I think I could comfortably switch to it in the future if I find the 4 AWG insufficient.  With 2 AWG it would mainly be just a weight penalty.]

For termination I'm using AMP uninsulated brazed-seam ring terminals (available from SteinAir and B&C, SteinAir being far less expensive).  Note that I've found that the best size match is to actually use a terminal designated 2 gauge sizes smaller than the cable, i.e. a 6 AWG terminal for 4 AWG cable, 8 AWG terminal for 6 AWG cable.

I'm crimping the terminals onto the cable using a hydraulic crimping tool from Harbor Freight (p/n 66150).  Here too, the best match between terminal size and tool die size is found by trial and error.  I've found that the 0 AWG die works well on the 6 AWG terminal, and 2 AWG die works well on the 8 AWG terminal.  Go figure.

I'm putting 1 inch of heat shrink tube over each terminal end, black on ground wires and red on (+) power wires.  1/2 inch shrink tube fits well over 4 AWG wire, and 3/8 inch shrink tube over 6 AWG wire.

"Hot" terminals will also get an MS25171 rubber boot to prevent accidental shorting by dropping a wrench, etc.  Size MS25171-3S over a 4 AWG wire, and -2S over a 6 AWG wire.

Copper bar

1/2 inch x 1/16 inch copper bar is used for short interconnects between the contactor relays, and also between the two fuse blocks.

Note that the cross section of this bar is therefore 1/32 inch squared, or about 40,000 circular mils, making it approximately equivalent to (just slightly less than) 4 AWG wire.  Van's drawing shows two of these bars stacked in parallel connecting the starter contactor to the master contactor, but I'm going with just one, as the greater resistance over this short length should be insignificant.  And if I did find that more was needed, I'd use a single 1/8" thick copper bar rather than stacking to 1/16" bars, which can trap moisture and promote corrosion.

Note that I had to slightly modify the distance between the holes in the copper bar to match the actual measured disance between the lugs on the contactor relays.  I also found that the lugs on the continuous duty relays (Cole Hersee 24115) the lugs are slightly angled down, which necessitated a slight bend in the copper bars to get them to sit flat against the nuts.  This was true of both 24115 contactors (master and ground power), so I guess it's not a defects... they just come this way?...

To protect against accidental shorting, I also put some shrink tube around the exposed sections of copper bar to provide insulation.

 

Back Up Next

 

 

Send mail to roee@kalinskyconsulting.com with questions or comments about this web site.
Copyright © 2003 Roee Kalinsky
Last modified: April 22, 2011

Visit my consulting web site at www.kalinskyconsulting.com for your engineering needs.