Yoke A/P disconnect resistor

[Estherville Aviation]

I have just completed a certified installation with two yoke A/P disconnect buttons. Followed the installation manual as closely as I was able. NOTE: Section 21.1 of the installation manual calls out installing a 4.7 - 5.3K resistor across the yoke disconnect button. If you install more than one yoke disconnect button, only install ONE resistor. DO NOT install a resistor across each yoke switch, it will draw the sense voltage too low and Skyview will never engage the servo(s). Lesson learned after hours of troubleshooting with tech support. Recommendation has been submitted to Dynon to clarify in manual.

 

[Rhino]

Pretty sure the resistor does not allow any ability to disconnect the autopilot. It's there to allow Skyview the ability to detect a broken circuit. Autopilot disconnect works fine without it, but an ability to detect a broken circuit is presumably required for certified aircraft, which is why the resistor is still optional for non-certified aircraft. The Skyview senses ground to detect a disconnect command. If you add two resistors in series, there probably won't be enough voltage drop off the second resistor to make that happen. Assuming this is the case here, and assuming you wired in series, the manual should have stated this better. It isn't 4.7 - 5.3k per switch. It's 4.7 - 5.3k for the whole circuit.

EDIT: And you might want to approach Dynon Support with this issue. If what I suspect is true, they may want to include it in the next manual revision.

Dynon Certified | Contact Us

Contact Dynon Avionics with questions, customer support or technical help about installing SkyView in your certified aircraft.

www.dynoncertified.com

 

[Estherville Aviation]

You are correct in the Skyview is checking to make sure the disconnect wire is not grounded. If more than one disconnect switch is installed, they are installed in parallel, and if each switch has a resistor, the resistance is lowered, getting below the 4.7K minimum needed.

 

[Rhino]

I meant parallel. They won't work in series. Sorry. Resistors in parallel will divide the current, but each one should still drop the total voltage, which makes your problem seem rather unusual. But, as long as you got it to work......

 

[Bill Putney]

I asked a similar question. I was hoping that the "Confidence" range was pretty narrow. I'd like to have an AP disconnect on each yoke with confidence checking for both. If the range was pretty narrow, each button could have a 10K resistor and if both were paralleled the resistance presented to the AP would be 5K. If there was a wiring fault to either switch, the 10K would not be recognized as a valid good circuit. It doesn't seem that Dynon has done that test or knows that answer, in any case they couldn't tell me for sure how it would behave.

 

[Rhino]

You may have to experiment. I think I recall seeing somewhere that the bias voltage is 5 volts, but I'm not certain of that. That might just be the EMS. It doesn't really matter how many individual resistors you put in each parallel leg, or what their value is. They'll each drop 5 volts, or whatever the total voltage is. The current could change dramatically, but the voltage in each leg won't. The real question is, exactly what does the Skyview sense, and how does it do it? There might even be an argument for pull down resistors. But unless Dynon is more forthcoming, it's a bit hard to say.

 

[IntAvTech]

I'm having trouble with this as we speak with just one switch and a single 4.7k resistor (Dynon-provided, at that!) across it. Despite metering the resistors I have and confirming that they're ~4.73k or so, when soldered in-circuit (and with nothing else connected) the resistance I see across the control wheel connector is 4.6k, which flags as "BROKEN", apparently. Not open or short but merely "BROKEN".
I would suggest that they should change the range they're looking for so that some common resistor falls close to the middle of the resistance range they're looking for instead of a range that puts two common resistor values at each end. Like, 4-5k or 5-6k instead of 4.7k to 5.3k. If they could also expose the raw value of the return instead of flagging both open and short as "BROKEN" it would also aid in troubleshooting. I'm assuming that my resistance is too low, but I gotta admit I'm guessing.
Currently adding a 300 in series to see if that sorts it.

Edit: 4.93k at the connector, .5 with the switch depressed and still no joy. Is this somehow configurable and I'm just not seeing the instruction?

 

[Rhino]

The certified manual calls for 5k, not 4.7. I'd try going up in resistance before going down.

 

[IntAvTech]

The certified manual calls for 5k, not 4.7. I'd try going up in resistance before going down.

Where? I can't find that anywhere. What I can find is this:
Revision O, Page 248 - "A 4.7 - 5.3k ohm resistor across the yoke-mounted AP DISC is required." That might imply a 5k but the two common resistor values in that range are actually 4.7k and 5.3k, at either end of the range they actually state.
Dynon provided a 4.7k. I first replaced it with one that actually read 4.7k (theirs read 4.6) and, when that changed nothing, added a 330 in series to increase it to (sadly) 4.93k, which shows how far I can trust the banding on my 330 ohm resistors.
Annoyingly, despite the caution message, the autopilot disconnect button works perfectly every time.
I suspect it has something to do with the trimamp. The APDC wasn't previously in the mix on the trimamp if you look at previous versions of the installation manual.

 

[Rhino]

In Revision M of the install manual, it's on page 246. 18.2.4, paragraph 2(B). Not sure if that's the latest version. Interestingly, it describes the button as optional, even though page 254 says it's mandatory.

The trimamp should have nothing to do with this. It connects to trim servos, not AP servos. The AP disconnect switch is connected to a serial Skyview network D9 connector for the autopilot servos. The trimamp is connected separately with a D15 connector. So they both connect to the autopilot, but they're electrically isolated, so they should in no way affect each other. At least not as far as resistance and voltage are concerned anyway. One is SV network, and the other is not.

 

[IntAvTech]

In Revision M of the install manual, it's on page 246. 18.2.4, paragraph 2(B). Not sure if that's the latest version. Interestingly, it describes the button as optional, even though page 254 says it's mandatory.

The trimamp should have nothing to do with this. It connects to trim servos, not AP servos. The AP disconnect switch is connected to a serial Skyview network D9 connector for the autopilot servos. The trimamp is connected separately with a D15 connector. So they both connect to the autopilot, but they're electrically isolated, so they should in no way affect each other. At least not as far as resistance and voltage are concerned anyway. One is SV network, and the other is not.

Check out the latest version of the wiring example - it shows the trimamp connected directly to the autopilot disconnect button on pin 6. (Page 10/15 and 11/15 on revision "G".)
I believe the intention is turning off the autotrim, in the event of runaway trim, with a single button press. Thankfully, I connected that line with a relatively accessible QD so it won't be a lot of trouble to test whether this is what's breaking the fault detection.

EDIT: Removing that line does, indeed, make the fault detection work properly.

 

[Rhino]

Without a Trimamp, the autopilot disconnect goes to pin 3 (yellow wire) of the DB9 connector for the roll and pitch servos connected to the autopilot. That’s actually the Skyview Network connector, which could go directly to the autopilot, or to a network hub. As you note, the Trimamp install instructions say to route the switch to the Trimamp, presumably to provide the function you mention. The Experimental install documentation has the same connections. Curiously, I can't find any mention of the Trimamp connection being in addition to the autopilot connection, or in lieu of it. Presumably it would need both to disconnect both trim and autopilot, as shown in the wiring diagram. Those two different paths could be introducing difficulty with the resistor installation. Assuming you connected the switch to both the autopilot and Trimamp, is the resistor installed before the wires break off into those diverging paths, or after? Depending on where it’s installed, the Trimamp connection might be preventing the resistor from providing the correct fault detection indication to the autopilot.

 

[IntAvTech]

The wires for all the autopilot disconnect circuits come together at the button as shown in Dynon's diagram. The button itself has a 4.9k resistance across the NO contacts and the other contacts go straight to a ground (I'm using both contact pairs on a 2PDT pushbutton for redundancy). Removing pin 6 on the trimamp made the problem go away and the fault detection is working, but I have no idea if autotrim will cancel with the A/P DC button now or not. I'm keeping it this way at the moment because the installation manual for the Dynon Certified said the fault detection was mandatory and the trim can be defeated by pulling the breaker anyway. Support is now aware of the problem.

 

[Rhino]

Auto-trim is run by the SV-AP-PANEL, so it should cancel. But that begs the question of why the disconnect needs to go to the Trimamp too. Hopefully Dynon will shed some light on that.

Is this another 182?

 

[Rhino]

This is what makes me wonder about this.

The autopilot and Trimamp will have a constant bias voltage applied, though I don't know what that specific voltage is. Without the resistor, they always see an open circuit unless the switch is activated, which is the same thing they'd see with a broken wire. The addition of the resistor adds a voltage drop across it, so it always has a voltage rather than an open. That's what allows a fault detection. If the wire breaks (opens), the voltage drop goes away, triggering the fault. The strange variable here is having both the AP and Trimamp connected together on the same wire. It's a pull down resistor, but what voltage it pulls down to depends on the internal voltage and circuitry of the AP and Trimamp. Could the bias voltage from the Trimamp be preventing fault detection by the AP by adding a voltage variable that takes the fault detection out of the desired range? I somewhat suspect they designed 4.7-5.3k to work with either the AP or the Trimamp alone, but that a different resistor value may be needed with both connected together. Since you have the quick connect to the Trimamp removed, you could attach a voltmeter to see what voltage the AP sees by itself without the Trimamp. You could then try disconnecting the line to the AP and see what the Trimamp gets by itself, and compare it to the voltage they see when they're both connected. That might lead you to a different resistor value that would work with both connected. Or you could just wait for Dynon to answer. I'm a tinkerer, so I'd be sorely tempted to experiment. But that's just me.

 

[Rhino]

Also, you said you're using both contact sets on the switch. You have them jumpered together with one resistor across both I assume, rather than a resistor across each contact set.

 

[IntAvTech]

Also, you said you're using both contact sets on the switch. You have them jumpered together with one resistor across both I assume, rather than a resistor across each contact set.

Correct. Not awake enough to do the unnecessary parallel resistance calculations.
Tempted to futz around with a variable resistor and find a range that works but don't have time. Also, without knowing whether the trimamp was ever intended to work with the fault detect...
Yes, this is a 182.