Avionics

As mentioned yesterday, with the main panel layout done, today I started thinking (again) about the layout of the stuff behind the panel. Also as mentioned yesterday, I wasn’t quite totally on board with the working layout I had after looking at all the actual components in place. So I started the day off just sort of staring and occasionally moving stuff around. I also got to thinking more about actual wire routing considerations, specifically the way the various D-sub connectors might affect things. To this end, I actually spent some time assembling a bunch of connectors and backshells so I could better visualize things.

In the end, even after moving stuff around, I was still dissatisfied. Everything seemed more cramped than I wanted it, especially the components on the baggage bulkhead. After lots more staring, I got to thinking about an idea I had a while back and, at the time, rejected – specifically, instead of placing the EMS and ARINC-429 boxes side-by side, figuring out a way to stack them. The thing with both of these boxes is that they have D-sub connectors on two sides, which means they needed decent real estate on either side for those harnesses. If I could stack them, that’d be less dead space I’d have to leave.

Now, there were two reasons I’d rejected this idea previously. First, I was concerned about the extra height of the stacked units – but that was several shelf iterations ago, and now I felt that was less of an issue. Second, I couldn’t visualize any stacking ideas that didn’t seem really unpleasant to deal with – most of them revolved around making a pair of z-profile brackets, and it seemed like a really fiddly thing to deal with behind the panel down the road. But today I had a better idea – thanks to cutting the panel (twice), I’ve got some scrap pieces of thick sheet lying around. Maybe I could fabricate something like the plate Dynon sells for stacking two ADAHRS units.

In the end, I decided to have a go at fabricating a mount, knowing that it might be wasted work if the packaging still didn’t work out. The idea was to cut a plate matching the base of the larger EMS unit, which would be mounted using the same stud mount I plan on the EMS unit, using spacers or standoffs to position ti above the EMS unit. Then I could just install nutplates on the mount plate, and screw the ARINC unit to it. Pretty straightforward work, really.

Cutting the plate was easy, as was getting the mount holes drilled for both the EMS and ARINC units. To help with potential interference with other stuff on the shelf, I positioned the ARINC unit off to one side (which, when mounted, means “higher up”) rather than centering it – this would provide more space near the shelf. Then I got my nutplates fitted and installed, and screwed the ARINC unit to the plate:

Next up were the standoffs. The mounting holes here are 3/16”, which makes this easy – I’ve got a bunch of aluminum tube that gets used for spacers on AN3 bolts like this all over the place. I even had a convenient donor piece: my first failed attempt at making the aileron trim pushrod. I worked out that the spacers needed to be 1 3/8” long to provide some space between the ARIND mount nutplates and the EMS unit, then did some cutting and finishing. And here we are, four nice little spacers:

Finally, it was time to put it all together! Only one problem: I didn’t have the right length of machine screws, since when I made my hardware order I wasn’t thinking of this application. But hey, I don’t need the actual hardware, I just need something to hold the assembly together so I can test-fit. I eventually found some AN3 bolts that would do the job, although they were way too long, thus requiring an absurd number of washers. Oh well, it gets the job done:

And so finally, it was time to go back and do the actual test-fitting…and hey, it works! I’d already decided earlier to move the CPI2 ECU over to the far right, mainly to provide better access to the fuses if needed. And that provides good room both fore-aft and up-down for the stacked units. It also lets me shift the voltage regulators back towards the center, which I like a little better…and maybe best of all, I can put the Skyview network hub over far right. That should make for some nice routing, since it’s 1) right beside the EMS and ARINC units, each of which have an SVN connection, and 2) convenient to the right HDX display, which is where I’ll start daisy-chaining the assorted panel units together, and 3) convenient to where the wire bundle going to the aft fuselage will exit the panel, which is good because I’ll have two SVN cables going that way from the hub (one for the roll servo, the other to tie into another network hub at the aft shelf).

I also rotated the CPI2 battery mount so its fuses face inboard as well, again for potential serviceability. The sole wrinkle here is that the Skyview backup batteries end up right between the ECU and battery for the CPI2 – this may make it tough to pull any of those fuses. But if that’s the case, I’ll always be able to remove the Skyview batteries if I need to – it’ll just be two plugs and four screws. Given that I don’t expect to be changing those fuses often, I think that’s acceptable.

This arrangement should also work well for potential servicing. The first step to working back here would be pulling the center panel unit – 12 mount screws to detach it, followed by seven plugs (one plug each for the two HDX displays, the GPS-175, the intercom, the radio head, and the AP panel, plus the SVN connector from the right-hand HDX display). I may consider trying to consolidate this into something like one or two cannon plugs or something, just for simplicity – I think that’ll depend on how all the harnesses end up tying together.

Anyway, back to servicing…with the center panel out, the entire upper shelf will also be removable, with about the same number of connectors: the two Skyview backup batteries, the D-sub and coax connector to the radio box, and the CPI2 connectors. The same thing applies here about maybe adding cannon plugs – that might be worthwhile to make this a little easier. I think the packaging for that might be more challenging, though.

Anyway, let’s get to some photos. First is a look at the area from the left side of the aircraft. On the baggage bulkhead, we have the two regulators closest, followed by the EMS/ARINC stack – with plenty of space for those connectors. On the shelf, we’ve got the radio box, CPI2 battery, Skyview batteries, and finally the CPI2 ECU:

And from the left side…the only thing to note here is the SVN hub, which wasn’t visible from the other side. You can also see a black cable coming from the CPI2 ECU and sort of just going down by the gear tower. This is the cable that will tie into the CPI2 panel controller, which will be on the right wing of the panel.

So there we have it – not a lot of progress in terms of tangible stuff today, but I’m starting to develop a solid mental picture of how everything up here is going to tie together. I really think the next steps here will be to get the panel wings cut and drilled for their components, so I can mock things up a bit more. I’ve also got probably another Spruce order to work up – I need the longer machine screws for the stacked EMS/ARINC mount, plus also maybe some hardware for mounting the CPI2 ECU.

There’s also one more component to add, a pretty late addition – I didn’t even think about having a carbon monoxide detector until this past week. These are pretty ubiquitous in piston aircraft, and most often are just these cheesy little cardboard things taped to the panel. The accuracy of those seems to be questioned, they have to be replaced periodically, it takes up space on the panel, and it relies on me noticing a color change.

This is, of course, all a setup for something better I found – an electronic CO detector, which can mount behind the panel, and is compatible with the Skyview EMS unit. What this means is that 1) I don’t have to take up panel space and 2) instead of having to watch for a color change, I can assign audible alarms via the EMS system – so if there’s ever a CO problem, I’ll get yelled at in the headset the moment it crosses a set threshold. The good news is, the detector unit is really small, just 2” square – easy to fit in back here.

So I think I’ve got the mission(s) for the week: more panel cutting, working up that hardware order, and firming up the harness layout behind the panel. Nice to have a solid path forward.

So tonight was about starting to figure out what the harness behind the panel ought to look like. It started with sitting inside and looking over my schematics, with the main goal being to decide whether to not I wanted to add in disconnects for certain modules (like the center panel or the upper shelf). After doing a bit of thinking and tallying up pins that would be needed, I decided that adding those disconnects wasn’t worth the effort. I think the main benefit of doing a single disconnect would be simplifying the service loop needed to allow removing the panel, but at the cost of more planning, more parts to order, and more wire terminations that could be a source of issues down the road.

With that decision made, it was time to start making a go at the layout. I decide to try and do some sort of iterative whiteboard work, sketching out the layout and visualizing the interconnects between devices. The idea is just to layout the paths that wire bundles will have to take, not to detail every single wire. The end result I’m going for is a diagram like the one I drew a while back for the harness to the tail of the aircraft – basically just laying out the wire runs and nodes, which I can then turn into measurements that become the baseline for harness building.

This, in turn, made me rethink some of my wire routing decisions. Initially I’d expect to have significant wire bundles exiting the upper shelf on both the left and right – on the right side would be power/ground connections, plus the beginnings of the run to the aft fuselage, and on the left would be the firewall forward runs, for the engine management system. The main hangup with the latter is that I intend to run the CPI2 FWF harnesses just inside the forward baggage well, to keep the wiring runs manageable. I think it makes more sense to run all the FWF EMS runs through that point, rather than splitting them up. That will leave me with just two firewall penetrations for wiring – the CPI/EMS one we’re talking about here, and one on the far left where the main power feed from the battery will enter.

I think this will be good for packaging behind the panel as well. On the left side I’ll have some SCAT tubing to feed the fresh air vent, which would potentially make routing a decent size wire bundle more interesting. By taking the bulk of the wiring off the left side, this makes things a little simpler. There will still be some wiring over here, but it’ll be limited to a pair of coax runs (for the radio and GPS antennas) and a few wires for fuel pump power and the left tank level sender.

Anyway, the end result of this was some fairly indistinct whiteboard material. Probably this entire diagram will go away and I’ll start over again, because while it decently depicts the interconnects between boxes, I don’t think it’s going to work for depicting the actual harness layout. I expect to end up with sort of two horizontal “trunks” – one that represents connections being router from the far right across assorted boxes, and another that links all the main panel stuff (and this can include the aforementioned service loop).

So at a glance it kind of seems like wasted work, but I think it did wonders for helping me get a better rental image of how things will go together here. I’ve just gotta do a lot more drawing…

It’s kind of challenging sometimes to keep a positive attitude right now. Even though I’m rationally aware that the stuff I’m doing is progress, emotionally it doesn’t seem super productive spending time standing around staring at things, grunting sporadically, and maybe sketching on a whiteboard from time to time. And that’s pretty much how I spent today, just really try to think about how to best route all the wiring runs behind the panel. Along the way, I changed my mind about some stuff…again. Sometimes I’m never quite sure if I’m being indecisive here and going in circles questioning myself, or actually refining things over time. I’m pretty sure it’s the latter, but the human brain sure is a pesky thing sometimes.

Anyway, the one big decision I made was about how to route the wiring runs going forward of the firewall. I’ve always been planning to bring the main power supply cables in through the left side of the aircraft, through a prepunched hole, but then there’s all the sensor connections and so forth. My working plan had been to route this bundle just inside the deep part of the forward baggage compartment – this would keep the wires from being exposed to potential damage in there.

But the more I considered the logistics of this, the less sense it made. Most notably, trying to work out the harness junction over by the right gear tower would be challenging, but also, I’d have to figure out how to get that bundle to turn a really tight corner to head for the firewall, which would mean also getting liberal with adel clamps or something to secure the bundle. The real kicker, though, was this – keeping the wires out of the baggage area to avoid damage seems good, but instead putting it right beside where my feet will go every single time I fly maybe isn’t such an improvement.

Now, there is a stiffener riveted to the side skin about halfway up from the floor, but it doesn’t stick out too far. Still seemed like this was a potential spot to route the bundle, and keep it protected and somewhat out of sight. Question was…how big would this bundle be? Kind of an important consideration. So I gathered up some of the preexisting harness pieces that would make up this bundle – the eight thermocouple wire pairs and the CPI2 bundles for the coil packs and crank trigger sensors. The only thing missing was for the other engine sensors…OK, time to go refer to my schematics. Turns out this is simpler than it seems, and all those sensors only require eight more wires. This was starting to seem manageable.

Finally, I checked the harness pieces I had against a snap bushing, and found one that they went through with plenty of room for a few more wires. Overall it looks like the bundle will be 1/2” to 3/4” in diameter. Very manageable for this spot. Really, the only thing to work out is how to secure the bundle. There are a couple of options here, but so far I’m leaning towards adding another piece of angle to create a protected space for the bundle…or I might just put some of my ample conduit supply here, just not sure the best way to secure it.

That can wait, because even if I don’t know the exact solution I’ll go with, I’m sure it’s a solvable problem, and all this head-scratching was just in service of figuring out where the harness would split back by the gear tower. Between this decision and a few other similar (but less complex) machinations, I finally got to a point where I felt comfortable expanding on my harness diagram, to cover the entire area behind the panel:

This lays out the general architecture back there: the harness will split somewhere near the right side of the upper shelf, with one bundle serving the components on the shelf, baggage bulkhead, and panel wings (the righthand branch), and another serving the removable center panel. That latter will need some kind of allowance for a service loop, so I can pull the center panel out and disconnect everything for, well, service. Of course, even as I write this, I realize that I forgot a few components here, but nothing that majorly changes things.

The one conclusion I’m coming to here is that it’s probably high time I stopped trying to envision the entire harness in my head, and started actually stringing wires and solving problems in the real world. Which means that I absolutely need to get off my butt and finish the fuselage riveting that I’ve been putting off for so long. Once that’s done, what I think I’ll do for the harness is build up the portions between the tail of the aircraft and the top of the right gear tower, and pull that through the fuselage. For the rest, including all the behind-the-panel and firewall-forward portions, I think I’ll be better off building the harnesses in place, rather than trying to carefully measure each one of those branches above and prepare harnesses on a table.

In the meantime, I can still go ahead and get components mounted to the shelves and baggage bulkhead, in addition to doing the cutting on the wings. But I still feel like I’ve got to stop putting off the riveting work. That’s sort of the elephant in the room that’s roadblocking real progress.

Well, at some point, it comes time to stop looking at devices mocked up on a shelf and actually start making some holes. I decided to make today that day for the upper shelf. This was mostly uneventful, just lining stuff up, drilling mount holes, and getting nutplates laid out. The one dumb mistake I made came while I was doing the mount holes for the Skyview backup battery mount – I’d unwisely left multiple drill bits lying around, and instead of grabbing the #21 for a #8 screw, I used the #12 for a #10 screw…so yeah, that’s now mounted extremely securely with four #10s.

With all the nutplates installed, it was time to add stiffeners. I’d been holding off on these until all the mount stuff was done, since they necessarily had to work around those nutplates. I decided to add two longitudinal stiffeners, roughly evenly spaced, plus another lateral stiffener right beside the CPI2 backup battery (the heaviest thing on this shelf by a good margin).

The stiffeners I made from some .032” sheet I bought a while back, using the cheap Harbor Freight bending brake I also bought a while back. This was a much nicer way to make this bent stuff than the hammer forming I’d been doing before. Then it was just time for more drilling, dimpling, and finally riveting. The finished product can be seen here:

Then I figured I ought to do a test mount of everything, you know, just to make sure it all fit up OK. Seems pretty decent:

Now I’ve just got to get this thing where it can be mounted in the fuselage. Just got to lay out a few more screw holes and get the matching holes and nutplates installed on the forward braces.

Well, in terms of structure, this thing is done. Tonight was just attach points laid out and wrapped up.

The shelf attaches to the two forward braces, both of which have a Z-shape profile. At the forward end, the horizontal part of the shelf lays on top of the middle of the brace, while at the aft end, the vertical part of the shelf sits just behind the brace flange. The forward spot was pretty easy; I laid out the holes, drilled them, and installed the nutplates. The attach hardware here will be AN3 bolts; way stronger than needed, but these mount holes will double as attach points for the adel clamps that will hold the wiring harness trunk.

The aft end was a bit more interesting. Since the shelf and the brace mate in a vertical plane, that means there’s no natural support for the weight of the shelf unless the screws or installed. My intent is for this shelf and all it components to be removed as a unit for service, which would probably make this a cumbersome choice. So I decided to add some support tabs on the shelf, to allow it to rest naturally in place even with no screws installed. Some scrap Z-bracket material from the fuel tanks worked great for this; I just made two small angle tabs and attached one at each end:

Then it was just the same routine of laying out the screw holes, drilling, and getting the nutplates installed. Here I’m just using #8 screws, which is probably about all I could manage anyway, as there’s a lot less material on the flange as opposed to the center of the brace.

So here it is, the shelf installed in the fuselage:

And a different angle, showing the forward attach hardware. I added the adel clamps just to give an idea of what the wiring routing would look like:

I’m kind of wondering if I should use screws here instead of the bolts. My thought with the bolts was that using a screwdriver deep behind the panel and up against the baggage bulkhead would be unpleasant. But getting these bolts started through the adel clamps isn’t exactly fun either. For now I guess I’ll stick with this, and see how much I hate my life when the time comes to install this shelf for real…