What's new
Van's Air Force

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

AP disconnect function problem

Tony_T

Well Known Member
Patron
Hi all,
I'm helping a friend with an AP disconnect problem. The autopilot will not stay engaged and gives a "disconnect switch locked" indication on the EFIS. We think the button itself is OK since we disconnected the molex under the panel and get the same indication. The button is supposed to ground a white wire to the servos to disengage them when things are working correctly. It appears there may be a short to ground somewhere. We are going to do some continuity checking and I would like to know where an ES-00019 connector shown on the elec. diagram is located. Does anyone know where this connector is located in the SV wire harness? It might be adjacent to the ES-00057 d-sub AP connector or even in it's backshell?

i-2dPjJ9Q-X2.png
 
Tony, based on the wire lengths shown on drawing that splice is 6” from the the AV-5000A Autopilot D-sub connector...(that’s also 24” from the 4-pin molex connector.)

I agree with you, The button is good and it’s probably a grounded wire, maybe even bad or miswired servo pulling the signal low,
 
Thanks guys.

Niner - I'm pretty sure an ES-00019 is a crimped splice.

Joe - Yep, I should have paid attention to those wire length call-outs. The splice should be easy to find and check.
 
More background & follow-up

It appears the autopilot on this -12 has never worked in the air. It would work on the ground but, not consistently when on the ground. Vibration when flying could therefore be an issue if there were an intermittent ground somewhere on the AP disconnect wiring or in the Dynon SV Classic or in the Van's control module (AV-50000A). The AP disconnect grounding is in the control module when the disconnect button is pressed. Lately, this SV has been booting up with the autopilot inop and a "AP DISCONNECT STUCK) warning. Once Skyview sees a stuck disconnect on boot-up it will not allow the AP to function.

We have checked the wire colors and pins in the fuselage harness to both servos and also the short harness and molex connector to each servo visually and with a Fluke multimeter. This all appears correct and we don't think there is a wiring error. We did find a couple of minor issues with the connector to the roll servo, corrosion and poor pin connection but fixing those did not solve the problem.

We kind of stumbled on a significant clue. When on a couple of times we had the AP running on the ground we found that tapping on the panel would cause weird AP indications and/or a stuck disconnect warning. Doing some detective work in the EFIS menus we found this:

i-M7s536t-L.jpg

Note the DISCONNECT SWITCH indicating PRESSED.

Then by just tapping on the panel next to the SV we see NOT PRESSED briefly for about a half second:
i-5HpNLTB-L.jpg


We tried another SV out of my dual side panel in the plane and did not get a AP disconnect warning and tapping on the panel did not cause a stuck disconnect warning. So our conclusion so far is that something is going on in the SV Classic in this ship, and it may have to go in to Dynon for evaluation. The Van's control module is not completely out of the woods as it mounts directly to the instrument shelf and tapping on the structure can affect it as well, but the SV is most suspect.

Since we are in the middle of good flying season here, the owner is considering options.
 
Last edited:
According to the schematic that I am looking at, the autopilot disconnect switch
is NOT connected to the Skyview. The wire from the disconnect switch
connects to an orange/yellow wire from the servos. I suggest that you
disconnect the org/yel wire from AV-50000A Optional pin 2 to see if the
problem goes away. If you get an error message about a faulty switch,
connect a 5K ohm resistor between the org/yel wire and ground.
I assume that the autopilot can be disconnected via the SkyView without a
need for a discrete switch.
 
Tony, I don’t think the issue is in the skyview hardware. The signal comes into the skyview through the servos via can bus. There is no discrete skyview input for this button. Both servos see the signal go low so the problem is likely on the wire side. That includes both servos, the wires and the van’s module.

Just thinking out loud here...Since your skyview works fine when installed here I wonder if there could be a configuration issue with his skyview when it learned the voltage on the disconnect signal input. Or, some how it changed the fault detecting mode from monitored, to unmonitored, and now the resister is pulling the signal down to a gray area between ON and OFF. (I don’t recall while troubleshooting if you measured an actual ground on the wire.)
 
Joe, I did take the ORN/YEL wire out of pin 2 of the OPTIONS d-sub, replaced the OPTIONS d-sub to the AV50000A, rebooted and the problem did not go away. So we deduced the wiring from that pin back to both servos pin 4 was OK. The SV booted up with the stuck disconnect switch warning.
The AP can be put into a standby mode by pushing the AP button once or completely off by pushing it twice. You can also turn off power to the servos by the autopilot switch on the switch/fuse panel. The Installation Manual says that the disconnect button must be installed because it has various functions, in the autopilot setup menus as for instance when calibrating the servos. Although I'm pretty sure once the autopilot setups are done that switch could be disconnected. We did disconnect it at the 4 pin molex under the panel and the problem did not go away on rebooting.
 
Tony, you misunderstand it. The signal goes from the button to the van’s module out pin 2 of the options plug to the servos. The servo sends it to the skyview via the can bus. Disconnect pin-4 at both servos and see what happens.
 
I did take the ORN/YEL wire out of pin 2 of the OPTIONS d-sub, replaced the OPTIONS d-sub to the AV50000A,
rebooted and the problem did not go away. So we deduced the wiring from that pin back to both servos pin 4 was OK.
Why did you deduce that? Unless I am misunderstanding you, I think the problem is with the orange/yellow wire.
Try disconnecting it at the servos to see what happens.
 
Tony, I don’t think the issue is in the skyview hardware. The signal comes into the skyview through the servos via can bus. There is no discrete skyview input for this button. Both servos see the signal go low so the problem is likely on the wire side. That includes both servos, the wires and the van’s module.

Just thinking out loud here...Since your skyview works fine when installed here I wonder if there could be a configuration issue with his skyview when it learned the voltage on the disconnect signal input. Or, some how it changed the fault detecting mode from monitored, to unmonitored, and now the resister is pulling the signal down to a gray area between ON and OFF. (I don’t recall while troubleshooting if you measured an actual ground on the wire.)

I don't know the logic in the SV regarding its ability to change modes monitored, to unmonitored. I do know that looking at the setting menu there is no option to set the AP disconnect switch to unmonitored. There is a option to have a broken AP disconnect wire warning provided a 5k resistor is on the wire, that is visible in the schematic picture and it's in the AV50000A control module. It may not be a choice if the resistor is there.

I did not get continuity when testing the ORN/YEL from Pin 2 of the OPTIONS d-sub to pin 4 on the fuselage harness molex to the pitch servo. Instead I saw about 144 ohms on that wire. I expected to see direct continuity. That wire was unusual. The disconnect wire was ORN/YEL back to the roll servo but there is a splice somewhere and the wire to the pitch servo is white with a number tag. Van's harnesses do substitute a white wire with a number tag sometimes instead of the color coded wire. I found the tag was tucked partly up into the molex connector and could not read the number. It could be problematic, but at one time my friend had the AP working on the ground and actually walked through the autopilot setups with the Dynon rep on the phone and everything checked out.

I did check both of the servos pin 4 to see if there was an internal ground in one of the servos. I saw no continuity to ground on pin 4 on either servo.

If the SV hardware is not the problem then I suspect the AV50000A control module since tapping on the panel or on the control module causes the servo disconnect status to change. Although we did not see that happen with my SV installed and I have to say my #2 SV is not configured or calibrated for his servos and by installing my SV we may not have gotten valid data. I feel I have to ignore that test. The only way to do that properly would be to install a SV, do all the network setup and calibrations on the servos, etc. as for a new install. That wasn't done.

We're kinda at a loss on this one.
 
Tony, you misunderstand it. The signal goes from the button to the van’s module out pin 2 of the options plug to the servos. The servo sends it to the skyview via the can bus. Disconnect pin-4 at both servos and see what happens.

OK, I didn't try that. I'll take another look at how that works. The pin 4 at the servos is those darn molex connectors and I have a couple of tools to remove molex pins but neither would work. I would have to cut the wire and didn't get that desperate yet.
 
Why did you deduce that? Unless I am misunderstanding you, I think the problem is with the orange/yellow wire.
Try disconnecting it at the servos to see what happens.

I'd have to cut the wire as I can't get the pins out of the molex connector.

I'm thinking replacing those molex with 9 pin d-subs would be an upgrade. I don't know if they have done that on the iS.

I am suspicious of the wire also, but that doesn't answer the question of why tapping on the panel causes the disconnect status to change.
 
I don’t know what rapping on the panel is doing, but it must somehow effect the disconnect signal so let’s follow that.
144 ohms sounds way too low. It could be a partial ground on a wire or a problem in one of the servos. Once both servos pin-4 are disconnected, check the ohmage of each servo and the wire back to options pin-2.

Ignore my comments about the software, I mistakenly assumed your skyview was functioning correctly when installed here.
 
With everything connected, pin 4 at the servos should have 5k ohms resistance to
ground. If the autopilot disconnect switch is pressed, the resistance should drop to zero.
 
With everything connected, pin 4 at the servos should have 5k ohms resistance to
ground. If the autopilot disconnect switch is pressed, the resistance should drop to zero.

Thanks, yes, I see that now. I didn't make that check.

I think the owner has put the inspection covers back on as he wants to fly the airplane tomorrow so I may not get on this again right away. I will pick up the thread again as soon as I can and get the opportunity to check that disconnect circuit on pin 4 of the servos per suggestions here.
In the meantime, I'm ordering the 9-pin D-sub parts and when I get back in there we'll change those molex connectors to D-sub. I especially now want to again check that pin 4 on the pitch servo that has the substitute wire.
I also got an email response from Dynon about the SV and they doubt very seriously, as you guys do also, that it is the SV that is the problem. You know, I could be chasing two problems here.

Thanks again for the help so far and I'll report back.
 
Here's the KAI drawing of the Garmin roll servo installation. Note the d-sub connector is integral to the servo case. I believe with the Dynon servos a short length of wires exit the servo case ending with a Molex connector.
 

Attachments

  • Screen Shot 2020-08-20 at 6.18.44 AM.jpg
    Screen Shot 2020-08-20 at 6.18.44 AM.jpg
    261.7 KB · Views: 175
Here's the Dynon autopilot servo installation drawing for the 12iS
 

Attachments

  • Screen Shot 2020-08-20 at 6.59.35 AM.jpg
    Screen Shot 2020-08-20 at 6.59.35 AM.jpg
    294.9 KB · Views: 174
Can you tell me what plan sheet shows the D-sub arrangement for one of the servos or a picture of the plan or picture of the installation?

Here’s the link to my autopilot KAI if you want to match the pin out.

https://www.dropbox.com/s/bck82dtj6sdt5kq/44AiS Autopilot.pdf?dl=0

Guess since you’re starting from scratch you could pin out any way you want, but if you follow the plans, you won’t have to worry about documenting separately your specific configuration.
 
Here's the Dynon autopilot servo installation drawing for the 12iS

Here’s the link to my autopilot KAI if you want to match the pin out.

https://www.dropbox.com/s/bck82dtj6sdt5kq/44AiS Autopilot.pdf?dl=0

Guess since you’re starting from scratch you could pin out any way you want, but if you follow the plans, you won’t have to worry about documenting separately your specific configuration.

Thanks guys, very helpful.
Regarding the pin outs, the molex to be replaced is a nine pin also so will use the same pin numbers and there should be no future confusion.

The iS build sure is nicely integrated, a lot of things like these connectors are just zip tied to something to secure them on the earlier builds.
 
Update - still confused

I don’t know what rapping on the panel is doing, but it must somehow effect the disconnect signal so let’s follow that.
144 ohms sounds way too low. It could be a partial ground on a wire or a problem in one of the servos. Once both servos pin-4 are disconnected, check the ohmage of each servo and the wire back to options pin-2.

Ignore my comments about the software, I mistakenly assumed your skyview was functioning correctly when installed here.

With everything connected, pin 4 at the servos should have 5k ohms resistance to
ground. If the autopilot disconnect switch is pressed, the resistance should drop to zero.

Stormy day today so on my computer instead of at the airport. Yesterday we booted up my friends SV on the ground and found the AP functioning. This follows our fix of the corrosion and possible poor connection on the roll servo molex.
The AP running in Expert mode in the status window, AP is green both servos are up, and in Roll Hold - Vertical Speed Zero:
i-qK7cmj2-L.jpg

We have the Disconnect Switch connector apart for this so the switch itself should not be a factor.

See the molex apart in this pic.
i-djwNtWd-L.jpg


By tapping on the panel or on the throttle we see the AP dropping off to standby, AP turns to white and servos are disconnected.
i-zsm3Vbm-L.jpg


I just don't see how it could be a wiring harness problem but I'm keeping an open mind. Right now we suspect the Control Module.
The Van's AV50000A Control Module contains the 5K resistor which purpose is to enable the SV to detect a broken wire on the disconnect. We do not see the "AP BROKEN DISCONNECT" and we should with the switch molex apart. So, perhaps the 5K resistor in the control module has an issue. I'm a little embarrassed to say we have not checked that and it would be easy to do at the disconnect 4-pin molex on pin 1 which should be 5K ohms to ground if the orange yellow wire is good and the servos are good.

I have checked both servos pin-4 and did not see an internal ground on the servos. I had the meter on buzzer and didn't check ohms.

With everything connected and if the orange yellow wire is good we should see the 5K on pin 1 to ground. Correct?

i-H3rGTKm-L.png


I may get a chance to check that tomorrow.
The inspection covers are back on right now, but we can check on pin 1 of the disconnect molex easily.

BTW, I talked with Dynon and they also do not think it is the SV.
 
Last edited:
I didn't read all the posts within this thread so I might be WAY off base, but I had a maddening AP problem that I finally found to be a connector loose in the MOLEX to the pitch servo (under the left seat).
Sure was thrilling when the AP servo setup/calibration in Skyview (dual HDX w/ knob panel) would finally complete correctly.
 
Tony,
If you are suspicious of the resistor or its circuitry, you can temporarily configure the Dynon to ignore it. That setting is buried somewhere in the AP settings menu. The fault sensing doesn’t do much good anyway, since the resistor its looking at is located in the 5000 box instead if being at the switch. Because of that, it does little to protect continuity to the disconnect switch.
Alex
 
Tony, let’s recap.

Look at the attached sketch and review the simplified signal path. The discrete signal input is located in the servos, not in Van’s box and not in the Skyview. The servos convert the discrete signal to CAN bus data and send it to the skyview. The only way to eliminate everything in the circuit is to disconnect the disable wire at both servos.


These are the important observations you’ve provided so far;
1) disconnecting the A/P button plug did not fix it
2) removing option plug pin-2 did not fix it
3) At some point you measured 144 ohms on the ORN/YEL wire to the servos
4) Tapping on the panel aggravates/resolves the problem

Based on those observations, the problem is not in the disconnect button, and it’s not in the van’s box. The problem is the 144 ohm reading on the wire to the servos. Based on your tapping observations, I would carefully inspect the ORN/YEL wire out of options plug pin-2. Look for chafing, pinched wires, insulation damage, missing snap bushings and the like. I would concentrate on the wire section in the avionics bay through that afterthought mess over the rudder pedals

sorry about the poor handwriting, too much effort to fire up cadd
 

Attachments

  • IMG_5627.jpg
    IMG_5627.jpg
    190.2 KB · Views: 180
That sketch helps a lot! So if there is a fault on the ORN/YEL wire between pin 2 of the options d-sub and pin's 4 on the servo, Van's box will report a disconnect to SV just as if the disconnect switch were pushed. I think I get it. My one quarter of electronics was many years ago and don't remember the concept of a can-bus. Had to look that up.

Concentrating on the ORN/YEL wire now, I'm trying to imagine where the short might be and influenced by rapping on the panel. Possibly where the wire bundle passes through the instrument shelf. Possibly at the wire bundle where it passes over the rudder pedals. I recall there is a screw that goes up thru the F-12125 Joe Gores wire bracket (sorry Joe). I have wondered many times if that unprotected screw could be chaffing on wires.

i-7WD4DNk-M.png


Then from there wire goes through various bulkheads to the servos. I doubt the short is located aft of the panel area since tapping the panel does not affect that part of the structure.

OK, I know where to check next, thanks much. Will report back.

P.S. The only way I can disconnect the disable wire at both servos is by replacing the molex's there with D-sub connectors. I can't get the pins out of the molex very easily. If I find chaffed wire somewhere on the panel area, and that fixes the problem, I won't need to do that.
 
Last edited:
old Uncle Occam would conjecture that it's connections that us amateurs make rather than the boxes & wire harnesses built by professionals
 
Tony, it’s not that complicated. The vans box is nothing more then a splice point. All the logic is in the Skyview.

Here is a sample fault detection circuit (resistor values are for demonstration only, I don’t know the actual values)


Looking at the circuit below, from a 5 volt supply, the input can see three possible voltages;
1) 0 volts when the button is pressed
2) 2.5 volts when the button is not pressed
3) 5 volts if the wire opens between the button and the input pin

Van’s missed the intent of the design here. The resistor needs to be at the button for proper detection. In the Van’s configuration, the wiring between the button and resistor is unmonitored.
 

Attachments

  • IMG_5628.jpg
    IMG_5628.jpg
    108.7 KB · Views: 166
Last edited:
Tony, I suggest that you bypass Van's AV-50000A. It is only a junction box. It
also has that resistor. But like Mike said, the resistor is supposed to be across
the switch, not in a junction box. That resistor is not needed anyway. If a wire
breaks, you will know it when the disconnect button doesn't work. There could
be something inside of Van's box shorting the disconnect circuit to ground.
Connect the white wire from the switch to the orange/yellow wire.
You are right about that screw. But you can not blame that on me. I used a
screw with a smooth shank and even gave Van's the part # (AN525-8R12).
They chose to use a different screw that has sharp threads contacting the wires.
 
Thanks guys,

I see a couple ways forward to get to the bottom of this one and am determined to solve it!
 
Hi Tony,

Don't know if this will be helpful for you or not ... but only takes a few seconds to try.

Early on when I first began testing of my SkyView system (which includes the optional AP panel) in my RV-12, I kept getting intermittent "AP disconnect broken" errors. This error was coming from the autopilot disconnect wire test option in the SkyView software.

I knew the disconnect was wired correctly and not broken. In addition, I also have AP disconnect switches wired on my grips and they worked just fine. Figured I have four ways to disconnect the AP ... switches on the stick, the standard red switch on the panel, the AP switch on the AP panel or the Autopilot power rocker switch on the panel.

As part of my run-up check, I verify the autopilot disconnect switches on the panel and stick grips work ..... so I just eliminated the the testing for the AP disconnect wire in the software by changing the Yes to NO.

Try changing the following: SETUP MENU > AUTOPILOT SETUP > DISENGAGE BUTTON > ENABLE BROKEN LINE DETECT from "YES"to “NO”. Not sure if turning this option OFF will help in your situation, but it won't hurt to try for testing purposes.

Troubling that tapping the panel will change your results ... but guessing the wire testing resistance may be marginal and any slight change in the wiring resistance is enough to bring in the error.

Your friend is very lucky to have you helping with this issue.

Hope this helps.
 
Last edited:
John, thanks for that suggestion. I have been considering that since the broken line detect is only marginally implemented here anyway. I'll set that to NO as you suggest.

The guy I'm helping is a retired dentist and is good with his hands. He wants to learn and is learning a lot. A multiple RV offender is helping along with myself. If someone without previous build experience buys a flying experimental airplane they are going to need a lot of help unless the airplane is perfect in all respects. This one had a few discrepancies that needed fixing. For instance, the canopy arms were not shaped according to plans and we had to remove the canopy to get the instrument cover off and work on the AP problem. And the canopy warning switch was not working -- fixed that. Then there is also another electrical problem with the audio fuse blowing that we haven't started to trace yet.
 
John, thanks for that suggestion. I have been considering that since the broken line detect is only marginally implemented here anyway. I'll set that to NO as you suggest.

The guy I'm helping is a retired dentist and is good with his hands. He wants to learn and is learning a lot. A multiple RV offender is helping along with myself. If someone without previous build experience buys a flying experimental airplane they are going to need a lot of help unless the airplane is perfect in all respects. This one had a few discrepancies that needed fixing. For instance, the canopy arms were not shaped according to plans and we had to remove the canopy to get the instrument cover off and work on the AP problem. And the canopy warning switch was not working -- fixed that. Then there is also another electrical problem with the audio fuse blowing that we haven't started to trace yet.

I can confirm and vouch for this /\ Fortunately, there's 3 other RV-12's besides mine, one of them has a hangar right next to mine. Two were builders, one other and myself bought their RV-12's second hand. Yes, you will learn a LOT, initially, about your plane, if a builder or AP assists you in going over your plane. But as they shape up and get closer to 100% with everything working, the joy of ownership and flying start putting that silly RV grin on your face, every time you come back from flying.

The plane will not stay perfect in all aspects, for ever, so you still need to put in the time and learn about your experimental plane, because there's few you can pay to work on it for you, at least around where my plane is based. There are, however a LOT of various model RV owners at the airport I am based out of.
 
Last edited:
Then there is also another electrical problem with the audio fuse blowing that we haven't started to trace yet.

Tony, the audio fuse issue may related to your A/P issue. Both are grounding intermittently and both have wires in the options plug. Audio power is ORN/BLK on pin-17.
 
Adding to what Mike said, since two wires are shorting to ground, maybe it is
where they pass through a hole in aluminum. Perhaps a problem with a
grommet or maybe that over rudder wireway. That would explain why tapping
on the panel affects the short circuit. Most electrical problems have a really
simple cause. The challenge is trying to find that simple cause.
 
Also check where connectors plug into the control box behind the panel.

I had an intermittent grounding of my trim motor and discovered a spot where the finger screws on either side of a plug into the box had a rough area to allow a good grip to secure the plug. Well this rough area wore through the wire and would short intermittently during the vibration of engine operation.

While checking on the ground I moved these wire bundles and plugs checking for anything and got the motor to activate. This lead me to the damaged wire (actually two right next to each other).

I did some reconfigurations of my wiring bundles and used small plastic tubing over these finger screws To cushion them from the wires.
 
Update -- Progress!

I had to start my ACI today and so we spent most of the day working on my airplane. But, before diving into that we removed the front tunnel cover to examine the area where the wire bundle goes over the rudder rudder assembly in my friends -12 with the AP problem (and the audio fuse blowing). We found this:
i-gThw2ZZ-L.jpg


Note the wire tray was never installed and the wire bundle is split into two bundles where it goes over the rudder assembly. I found by moving the wire bundles the AP works and no amount of tapping or even hard rapping on the panel results in the AP dropping out like before.

So I think we know about where this problem is and what it is. As Mike and Joe advised it is not the SV, not the Van's Control module and is a short to ground on two wires. And, the Audio fuse also is not blowing!

It's not fixed though. We still need to find the place where the chaffing is taking place and then how to protect the wire bundle. I'm thinking we should also install the wire tray or trays with Joe's smooth shank screws. We stopped at this point to work all day on my ACI but will get back with a report in a few days when we have time to work on it.
 
I found the builder hadn’t installed the tray in my aircraft. I bought one and then decided it looked like a difficult job for an old bloke, and also had concerns about poking a screw up through the bundle. Instead I cut a piece of rubber about 3/16” thick and put it over the rudder tubes. It hasn’t shifted in 400 hours. The counter rotation of the tube resists movement. The fore and aft edges are chamfered to provide a smooth path for the bundle.

Jack
 
I fabricated a second tray and split the wire bundle to help ease the congestion. It’s an awkward area to access.
 
Problem located!

Yep, it was simple and obvious. You can see in this picture the ORN/YEL and ORN/BLK wires were chaffing on the cabin heat cable where they crossed up under the panel. The ORN/YEL is the servo disconnect wire and ORN/BLK is audio power.

i-gkrGrxF-X2.jpg


We sure went round and round on this suspecting wiring at the servo connectors, the SV, the Vans box, etc., geez. We did find problems at the roll servo that we fixed so there could have been more than one issue with the AP.

We didn't connect that the audio fuse blowing was related but Mike had that right on. Test flight is still to come, but we think this is it.

This forum was really helpful. Thanks.
 
Back
Top