Trim conflicts with AP Alt Hold ?

[Guest]

Hi all,(this questions was also put in the home cockpit forum, but without results yet..)i'm building a little cockpit addon that gives me some extra axis' (3) and one of those is going to be the elevator trim wheel. In testing that function within FS i find that the AP ALT hold doesn't function right anymore. Because the AP doesn't have control over the trim wheel and it receives constant (and wrong) input from that axis. At least that's what i think happens. Does someone know if this is right and is there something i can do about it ?If there is no cure for this, does anybody know how a unit like goflights gear/flaps/trim combi handles this ?Thanx for your help,Cas

[rcbarend 37]

Hi Cas,I think your analysis is correct.I think pitch and pitch trim are handled different by FS2K2 than e.g. the throttle axis.When the AP SpeedHold is enabled, you can move the throttle lever anyway you want but FS2K2 will not respond to.Unlike pitch. When AltHold is on, and you move the elevator, FS immediately responds with changing the pitch trim.But, when it also receives trim input from an axis, I think it gets "confused".The reason why other devices than yours may not have that problem, is probably caused by how FS2K2 deals with Controller inputs: FS is trigger-driven, i.o.w. it only sees/reacts on an axis in case the value changes.And therefore there is a big difference between an axis connected via the gameport or an axis on USB (including a potmeter via a gameport-USB convertor).Reason: any axis suffers from "jitter" (small variations in value, caused by EMC). This means that an axis connected on the gameport gives a continouous "change" recognised by FS.However, axis connected via USB give a steady value when not moved, since the axis is sampled into discrete values. And if the value changes due to "jitter" falls within the resolution of the sampling, your PC doesnot see any change.You can easily verify this effect with the throttles (provided it is well calibrated).- Set the throttles to idle.- Press F3 to give some throttle, the panel lever will move.- Now move the throttle lever a bit.- FS now sets the panel level to the value of the axis.- If you have a gameport throttle: - Press F4 for full trottle - You notice that the panel lever immediately falls back. (due to the jitter)- If you have a USB connected throttle: - Press F4 for full throttle - The panel lever will now stay at Full T. untill you move the throttle lever a bit.This is also the reason why some people have problem with Reverse Thrust (via F2) not holding: even with the external throttle at idle, "jitter" may cause FS to detect a non-idle value of the axis.Now, how can you prevent this ?Since you have this problem, I assume you have the Trim axis via the gameport.Solution:1. Define a null(dead-)zone for the trimaxis, and make sure that your trim wheel is in the nullzone when you activate the AP so FS doesn't see jitter changes anymore.2. Disable the controllers while enabling the AP.E.g.Using the "buttons" facility of FSUIPC, define a button that toggles the AP master switch AND toggles the Controllers on/off.So, when AP is enabled, the controllers are disabled.Ehmm... not a good solution: how would you enable the controllers again with a controller button is the controllers are disabled :-)..What you really want is to disable only the Trim axis (or set the Sensitivity to Min via a controller button). But I doubt whether you can enable/disable individual axis via the FSUIPC "buttons" functions.The first solution will work anyway.Cheers, Rob

[Guest]

Rob,thanx for your great reply! It has helped me to understand my "problem" a lot better. For once, i didn't know there was a real difference in the way USB corresponds compared to a gameport. However, my testing sofar was done through the mixture axis on my CH Yoke (and thus a USB device) and the real thing will be done through a hacked USB joystick (MS Sidewinder). So there is really no gameport issue for me. I haven't experienced jitter on those USB devices at all (i have on my gameport propedals connected through an adapter, but that's another story!).Anyway the tests on the yoke didn't do it for me, and after your explanation i wonder why. Tonight i will do some more testing on one of the axis'on the yoke and also on the Sidewinder.I also thought about disabling the trim axis through FSUIPC when i fly AP, but i haven't found a function for that in the buttons section yet. Will look into that tonight too.Again thanx and i will report about my findings here.Cas.

[rcbarend 37]

Cas,The best way of proving that the axis has no jitter, is to look at the INPUT value for that axis in FSUIPC.If that stays stationary all the time, or if the deadzone trick doesn't work, then my assumption that ALL FS axis are trigger-based (like I described for throttle) is obviously not true. But I doubt that.Also, make sure that no other controller axis is driving the TrimPitch function, since in FS it is possible to assign two axis of two different controllers to the same Axis function. Which obviously given unpredictable results.Rob BarendregtPS. Saw your name in the other forum; (van der .... ) sounds like we're both Dutch :-).

[Guest]

Cas, I too tried to use my mixture axis on my CH yoke as an elevator trim axis and found the same thing you found. I finally gave up and assigned it to flaps Via FSUIPC and went back to trim buttons on the yoke. I too wonder how Go Flight achieves this with their trim wheel. I would love to build a large trim wheel for my cockpit and have it drive a hacked out pot from my CH yoke lever via a gear reduction of some sort but am afraid to do so because of the interference with the autopilot trim. If you manage to get this to work would you pleas post the solution here. Thanks,http://www.ktone.org/images/FSD_ken.jpg

[rcbarend 37]

Cas, "Blow",I just tried it myself, and it works exactly as I described.What I did:- I assigned the Throttle level of my CH Yoke LE USB to TrimPitch, with Sensitivity Max and Nullzone Min.- I calibrated TrimPitch with FSUIPC: a small deadzone at both exterme positions, and a nullzone at center position. This guarantees that in these three positions I don't have any pitch variations.Test1: - Set the throttle lever (=trimpitch) to centre.- Take off in the default 737- Climb on the AP to 5000 feet, where it nicely leveled off.- Increased altitude to 7000 feet.- 737 climbed to 7000 and nicely level off again.So it works !!- Moved the Pitch trim, and now the AP was confused :-)Test2:- Set the throttle lever (=trimpitch) to full pitchup- Take off again (needed to push the yoke to make sufficient speed)- Climbed out again to 5000 feet, slowly letting the Yoke go to center.- And again, the PitchTrim nicely corrected and leveled off at 5000 ft, with elevator and pitchtrim nearly centered.Conclusion: the PitchTrim axis can remain connected on AP, if only you make sure that it gives no changes while AltHold enabled. So putting it in a deadzone when you activate the AP solves your problem.And I can't imagine that this doesn't work for the Prop or Mix lever on your Yoke: for FS2k2 it's just an axis....And a last remark:I could hardly find any position of the lever outside the deadzone, where I didn't get any fluctuation between two values on the FSUIPC Input. Which proves my point again: you're problem is caused by FS seeing changes in the axis value even if you don't touch the lever.. And one detected change is enough to mess up the trim at level flight.Rob

[rcbarend 37]

"Blow",See my post below.If you would connect your trimwheel potmeter to a gameport-USB convertor instead of your hacked USB joystick, it would certainly work.Reason: the resolution of these cheap convertor is much less than your USB joystick; most convertors only have 32 position over the full range, so any noise (jitter) will not result in an output change.Or set your trimwheel in a center-deadzone before activating the AP.Rob

[Guest]

Rob (yes we both are dutch) and Blow,nice to see i started such an interesting discussion.Sadly i didn't have time yesterday to do those tests myself. But i completely rely on your tests Rob. My conclusions on this:1. A USB axis does produce jitters so without changes isn't suitable for trim (combined with alt-hold off course).2. In a normal lever situation (as on the CH yoke) with a visible lever a deadzone in the middle is possible. In my intended situation however (with a trimwheel that makes about 5 full rotations through some gearwheels) that won't work without any indication of the trimsetting on screen. And my A/C don't offer that all. So i decided to go on with my project (materials where delivered yesterday anyway) and concentrate on trying to find a way to disable elevator trim when using alt hold. If that doesn't work i won't use althold at all, as all my flying is done on small A/C (biggest one is the beech G18) and it's more fun anyway to have something to do when flying.....My time last night was spend on building the mechanics for my new cockpit addon. Doing that with some old Meccano parts i had lying around. Finally managed to support 3 axis and still fit it in the case i ordered for this. This weekend it's time for some serious drilling and assembly of it all. When i have it all working (there are also 20 swithces/buttons involved in this case) i will post my experiences. One question remains however: How does goflight get this done ? Anyone out there that uses that module (with combined gear/flaps/trim functions) and can share his experience with regard to alt-hold ????Thanx for your input guys, and i will be back.O yeah, Rob: wordt het geen tijd dat Sinterklaas jouw LE stuurtje vervangt door een uitgebreide ?? Of heb je ook die functies al zelf gebouwd ?Cheers,Cas

[rcbarend 37]

Cas,Do a search on author "Barendregt" in the library, and you'll see why I don't need another Yoke :-)With some more projects under way ......Cheers, Rob

[Guest]

Rob, "Blow", and others interested,this weekend i had time to build my home cockpit thingy, and late at night on sunday i did some testing on FS. I like to share my experiences in the home cockpit forum, which is more appropriate anyway. So hope to see you all over there.Cheers,Cas