Bonanza Autopilot and Aspen combo problem?

gkainz

gkainz

Final Approach
Joined
Feb 23, 2005
Messages
8,401
Location
Arvada, CO
Display Name
Display name:
Greg Kainz
for discussion or ideas ... I'm posting the following message from our club maintenance officer regarding an issue with our '87 F33A Bonanza, Aspen Pro 1000, Garmin GTN750, KCF 150 autopilot.
-------------------------------------------------------------------
"Gentlemen: we have two continuing autopilot problems, in spite of frequent avionics shop visits, adjustments, and multiple component changes: Porpoising (now identified as pitch pumping) and pitch-ups at autopilot disconnect.

Most of this info comes from a long conversation with a Senior Tech at a major repair station, and from the owner of “our” shop.

Our autopilot is a KFC 150. It uses 3 servos: One for roll, Two for pitch. We aren’t having any known roll servo problems.

Pitch: A PITCH servo drives the elevator. A pitch TRIM servo drives the trim tab.

When WE fly the plane, we FLY THE ELEVATOR, and then trim as needed. This is NOT frequent, and we never use hundreds or thousands of rapid up/down reversals.

When AUTO flies the plane, the computer (the panel mounted controller) will ask the PITCH servo to move the elevator. Resulting cable tautness signals the TRIM servo to neutralize the cable pressure.

The MAIN part of the KFC 150 test cycle is a computer check (four red light flashes). First flash: pitch signal sent. Second flash: Signal returned (OK). The next two flashes mean another signal was sent and received, all ok.

Right now, this test in our Bonanza is normal, which means the computer is ok. ( It was rebuilt recently)

Our Aspen Pro 1000 uses an electronics box (the EA 100) to monitor pitch and roll, and send emulated pitch signals to the computer. Big Glitch: Computer needs adjustments to correctly read EA 100 signals. This is not a ‘once and done’ thing. If ANYTHING is changed, like installing a new component, or loading new Aspen software, another calibration is needed to assure the computer can read the artificial pitch/roll information correctly. This is called an “AutoPilot Alignment”. We had this done last month. Our satisfaction level was about 75%. We went back for a software update - but did NOT get a fresh alignment. Satisfaction level dropped to near zero.

Recently Learned: We have pitch pumping, not porpoising. 1 second intervals. This is evidence of mis-alignment with the EA 100. MY THEORY is that “pitch pumping” is seen when the pitch servo responds EVERY second to erroneous pitch commands: Up, down, up, down, etc. Servos are built for NORMAL ops, and even turbulence does not exercise them rapidly. I think “pitch pumping” destroys servos. It also makes using an affected autopilot an intolerable experience.

Our second problem: Severe nose up pitch: There is another servo ground test. Here’s the sequence, and the report of our servo action: Push forward gently, servo reacts, trim servo follows up, starts trimming nose up. FAIL. Nose up trimming is supposed to stop when back pressure is released. FAIL. It does not stop. It is probably that inflight conditions have occurred that caused the nose up trim to begin. When it has not stopped, the next disengagement was violent.

Continuing: Gentle Back Pressure: Trim should start nose down. FAIL. It does not. There are separate switches for up/down. Imagine the rapid solenoid action and motor reversals with one second intervals. There is also a tach generator that sends rotational speed signals. Imagine reversing it, same intervals. And there is a tang that can break. Don’t know anything else about it.

I have a received another rebuilt pitch servo. I will install it soon, but this probably means is that we will start destroying it if we let ANY pitch pumping happen. There is no time in the schedule right now for a trip for alignment. You can TRY the A/P, but punch it off at the first sign of pumping, even if it is a small magnitude pumping action.

MY OPINION: If we can’t cure the pumping, we should find a REAL attitude indicator to drive the computer— which was NOT BUILT for “emulated signals”. It will take big bucks, but our experience with our other aircraft (real attitude indicators) was 100% satisfaction. I think that is where we should go.

I’ve given up any hope that we will ever be happy with the EA 100 electronic gyro, and its “emulations”."
-------------------------------------------------------------------
 
Seems like a note to whoever makes the EA100 to see if there's a way to have it give a bit of hysteresis before sending a change signal would be a worthwhile communication, along with a copy of this letter. Essentially what he's saying is that it's "too accurate" and triggers too much movement and response of the servos. The "real gyro" version would operate slower.
 
Reference to EA1000 is in fact the Aspen Evolution 1000 (don't know why Aspen Evolution - AE, becomes EA but it does). I believe we've had conversations with Aspen, including recommendations from them to get some software updates and patches. Done. No change.

Snip of correspondence from Aspen regarding our request for assistance

(Your shop's technician and your club's maintenance officer)... "have been brought up to speed on a set of software tools that were released Spring 2015 which allow some extended adjustments of the EA100 settings which correct these types of pitch oscillations that you are experiencing. These tools have been used on aircraft that have experienced pitch oscillations with good results."

Fix applied with no resolution.
 
Last edited:
Ahh. I remember now when you posted that letter. Sounds like that's the "alignment" tool/software.

I saw where his note that there's no openings in the schedule to do the alignment, but then he also said they're required at any software update and they did a software update? Huh?

Right now the wording of the maintenance officer still sounds like the theory of what's happening is technically unproven. It should be measurable, like anything should.

But they'd probably want the alignment procedure done first by someone they trust.

Looking at various docs on the web, the original system included a gyro that output an AC signal that changes by 50mV (and perhaps also changes phase) to indicate a specific number of degrees of change in pitch.

The King system has both the panel display unit and a "computer" which takes that analog signal from the gyro and turns it into flight guidance and servo control voltages.

The whole system relies on that "computer" measuring voltage changes and then reacting to them, from the gyro, a "tach" voltage from one of the servos, and all sorts of lovely 1980 vintage analog goodness. Haha.

And also supposedly keeping things within some hard limits for total pitch movement and total trim movement.

So all Aspen did was come up with a way to pretend they're the gyro and sending analog voltage changes to the King "computer".

The Aspen MFD drives an additional box they make called an EA100 that takes the data from the AHRS and emulates the old gyro.

All that thing has to do is take digital pitch info from the AHRS/MFD and turn it into an analog voltage. It's a pretty simple box.

But...

Google turns up multiple stories of the two together -- the EA100 and the King "computer" doing a slow pitch oscillation continuously when it happens but not consistently.

Interestingly there's plenty of stories of airplanes doing it with the original King gyro as it started to fail, too.

And worse, there's plenty of stories that the culprit was the king "computer" and stories of the servos sending back bad feedback on their position sensors and/or the tach sensor so that the computer fights against these bad indications from all directions in the circuit diagram.

I see not much "good news" from people who have fought this on both the 150 and 200 versions of that King AP.

Apparently the proper way to measure and test all of this is a King test set that probably very few shops had and nowadays probably fewer. One search turned up a guy on Beechtalk who said he had one and would hook it up and look at things for someone if they flew their plane to him.

A few of the posts also mention that the King "computer" needs alignment besides the EA100 needing one. Or maybe that's what the EA100 alignment is really doing, adjusting analog voltages to the ranges that work right with a particular King "computer".

Seems like the easiest thing to do would be to disconnect this EA100 box from the AP and hook up a cheap digital capture scope to the fake "gyro" output pins and go fly the plane straight and level and see if the "fake gyro" output is bumping up and down on the pitch pin, voltage wise.

If it's not, that would break the problem in half and say that the real problem is over in the AP somewhere.

Also a really dumb question but worth asking. With an oscillation that slow has anyone checked to see if it's oscillating in time with the beacon?

I doubt it since there's all sorts of stories of this pitch oscillation being a common failure mode of the King AP even when configure with the original gyro, as it starts to fail or experiences a vacuum system leak.

Messy. And too many feedback loops that could trigger the pitch issue.

Seeing if the output of the EA100 is actually bobbling up and down voltage wise, seems Ike a pretty good start.
 
I doubt this is your problem, but my autopilot can get its left and right from the IFR GPS or the VOR. When running off the GPS things work MUCH better when the GPS is set to its most sensitive. I think its 0.3 per tick mark. It also seems to work better the fasterthe plane goes.

So maybe there is a sensitivity setting?
 
coloradobluesky said:
I doubt this is your problem, but my autopilot can get its left and right from the IFR GPS or the VOR. When running off the GPS things work MUCH better when the GPS is set to its most sensitive. I think its 0.3 per tick mark. It also seems to work better the fasterthe plane goes.

So maybe there is a sensitivity setting?
Click to expand...

Ha that's interesting because I ran across that one doing the mass Google search last night too. If the GPS is a Garmin, there's a software fix for that. It's not supposed to tie the Airinc interface accuracy to the screen display accuracy but it did. They fixed it.
 
Hey Nate - thanks for the thoughts and feedback. I haven't had time to read in depth yet, but to your first point, the timeline confusion is really muddied by numerous trips to the avionics shop, multiple troubleshooting exercises with the AP (replaced with rebuilt AP, swapped in known good loaner) and other maintenance actions that I don't have at hand right now. And, yes, at least 1 trip into the shop for the alignment recommended by Aspen. I believe the MO is referring to yet another appointment we have for further troubleshooting when he said "no openings for alignment". And I just saw an email from him yesterday where he noted "the whole avionics shop is down with flu, so our appointment is delayed".

I'll have to get the diagrams and docs you found and review them if I ever can manage to shave off a few spare hours somewhere... sometime.
 
I had an interesting time getting an Aspen to correctly talk to an analogue Century autopilot. You really want to go to a shop which has solved many of these problems. In my case the analogue converter on the Aspen had to be matched to the autopilot to solve all the problems. The shop has to have enough converters on the shelf to find the match or you wait for units to be shipped back and forth.

My gut feel here (and that's all it is) is that the digital side of the Aspen/EA100 is fine and there is a compatibility problem matching the Aspen analogue output signal to the autopilot. First get the damn thing aligned and if that doesn't solve it make sure the units are matched to each other.
 
As others have discussed above, ensure the Avionics shop has a KTS 158 test set. This is required to follow the manufacturers adjustment and return to service of the KFC 150.
 
Folks, after 14 months of mating my A36 Bonanza with a KFC-150 to an Aspen Pro 1000 PFD, I'm pumped out. I'm not sure the pumping has been solved, but it's taken its toll on my patience and payouts to try to resolve the issue. At first the 'pumping' happened all the time the AP was engaged, but after many trips to the shop, software 'tools' and thousands of dollars, it may only happen on descents now. Or it may really be resolved. We'll see. It behaved on the first low level flight home, as it has before.

Aspen seems to be aware of the problem, especially in Bonanzas, and has helped with ideas, tech support and conversations with me and the shop who's now willing to work with me to solve this. I'm very appreciative for the shop's willingness to work with me on this matter. (Not same as install shop.)

I've received NO financial help from Aspen. Only thoughts and prayers, along with best wishes, sympathy and awareness of the problem. They did send an email inviting me to upgrade to the MAX, just last week. It hit my inbox the plane was..wait for it...in the shop trying to solve the pumping. I suggested a nice gesture might have been a discount on the MAX upgrade. Same answer I got for some financial help getting the pumping to stop: 'NO.'

When I removed to old as equipped Bendix King system it was working just fine. Wish I hadn't, and just kept flying the old system. I could overhaul the gyros for less, I am sure.

I do like the Aspen, don't get me wrong. I just wish it worked as advertised, in my plane.

I know this thread is old, but I thought as a late bloomer to the Aspen (since a new one is now almost ready to ship), there may be some pilots still pumping along, or am I alone?

Eddie
 
flyingcheesehead said:
Yikes. One more reason to go Garmin. :eek:
Click to expand...

Except, Garmin has nothing remotely equivalent to the Aspen for autopilot guidance.

In any event, I really like Nate’s troubleshooting protocol, and agree: until the discrete elements are tested, independently, we can’t know what the cause is.
 
SCCutler said:
Except, Garmin has nothing remotely equivalent to the Aspen for autopilot guidance.
Click to expand...

Huh? Sure they do. It's called the GAD43e, and it takes the same place in a Garmin glass system as the EA100 does in the Aspen system.
 
We may fixed! After 15 months, a flight well behaved. A couple of more and I'll be convinced. Involved putting the plane on jacks, leveling and calibrating-carefully.
So, who's up for the upgrade?
 
Seems to be working. Another flight today, IFR today, and I believe a careful leveling and calibration did it. That was quite a workout of my patience and wallet. But it sure is one heck of a product, I must say. Choose your installer carefully. I didn't take the lowest bid, by the way.
 
Are there any updates on this?
I'm having an intermittent problem with the EA-100 disconnecting and also the pumping.

What shop did you use?
Does anyone have any information on the - Tech Note 2010-10
This is supposed to be used to align the EA-100?

Thanks'
Dave
 
If there’s any interest in updating this thread, I had a very similar divergent porpoise that I believe was solved with proper Aspen KFC150 calibration. I’d be happy to go into more detail for those interested.
 
Not sure if it's been mentioned before already, if so apologies, but the ASPEN PRO EFD1000, when set to "gpss" mode will just pass through the GPS steering commands received from the G430 or whatever you have. This introduces major lag, and it will overshoot corners (by quite a bit). This is exacerbated when running
coupled (rnav) approaches.

I have the ability to switch my Autopilot between the Aspen and the garmin. And to be honest the only time when it is on the "aspen" side is to sync the baro to the AP.

Since the aspen doesnt have a GPS input natively by itself, if you have the ability let the AP take the commands directly from the GPS (takes one messenger out of the sequence too).
 
You must log in or register to reply here.
Back
Top