[NSRCA-discussion] Twin elevator servo issue using Futaba 9 ZAP

Sat Aug 12 06:13:36 AKDT 2006

Hi Peter,

  You previously asked if I was using Ailevator function, I am.

  I'm suspicious that the matchbox improved things that much.  I  found I needed to set my Ailevator function percentages to exactky 100%  or the intermediate points would be way off.  Don't use Ailevator  ppercentages to match end points.  I set my Ailevator percentages  to 100%, used STM to match the center points and ATV to match  endpoints.  That got me pretty close throughout the entire  range.  Then I used a PMX to clean up the rest.  The PMX is  set to 0 at the end points and center, and up to 5% at intermediate  points.

  Of course this is only static and doesn't help the dynamic delay  problem Brian was talking about, but I don't think that delay is much  of an issue unless you really yank the stick.

  Ed

Peter Pennisi <pentagon.systems at bigpond.com> wrote:              v\:* {behavior:url(#default#VML);} o\:* {behavior:url(#default#VML);} w\:* {behavior:url(#default#VML);} .shape {behavior:url(#default#VML);}                    Hi all,

    I appreciate all the feedback in relation  to this topic.

    I have had a closer look at my setup and  have come to the following conclusion.

   The       mechanical setup of both elevators is close to identical but not exact.       This is certainly part of the problem.   
   Servos       are not identical and therefore contribute to the problem    
   There       is a slight delay between master and slave channels if using the Ailevators       function  

    I went out and picked up one of the Futaba  MSA-10 servo matchbox. This item certainly removed a majority of my mismatch without  touching too much else. I had originally started to use a point mix and tried  to fix it electronically in the radio but it all became too difficult.

    I went out this afternoon and flew the  model with the matchbox and I could certainly tell the difference. It was a big  improvement over what I had before; believe me I had spent a considerable  amount of time trying to get it right. The model tracked perfectly in very  tight to medium sided loops both upright and inverted. I also noticed a big  improvement in the hourglass maneuver as the model tracked more precisely.

    Peter 

---------------------------------

    From: nsrca-discussion-bounces at lists.nsrca.org  [mailto:nsrca-discussion-bounces at lists.nsrca.org] On Behalf Of Bob Richards
  Sent: Friday, 11 August 2006 8:31  PM
  To: NSRCA Mailing List
  Subject: Re: [NSRCA-discussion]  Twin elevator servo issue using Futaba 9 ZAP

        Brian,

        I think there are two problems. 1) Movement commands from the radio  arriving at the servos at the same time (dynamic) and 2) Control surface throws  matching through the entire range (static).

        What you are describing is a dynamic problem, ie: does the radio do all  the mixing and send the information to the servos *at the same time*.  In  other words, while the servos are moving, one is lagging behind the other since  the radio is processing one channel before the other. Also adding to the  dynamic problem are the speeds of the servos, they must be the same.

        The other problem is what I was talking about, when the servos stop,  are the control surfaces tracking together. This depends on the linkage geometry  as well as the servos being linear.

        A matchbox attempts to correct both problems. However, a matchbox only  matches 3 points in the entire travel range, and the problem described (as I  understood it) is that the throws don't match halfway between the neutral and  endpoints.  It would be nice if there were multiple points in the range  (5, or 7) that could be programmed in the matchbox. This feature could even be  used to program in expo, seperate from the transmitter.

        Servos can have linearity problems due to the manufacturing tolerances  of the feedback potentiometers. I have seen pots (not from servos) that were  "adjusted" by milling away material around the perimeter of the  conductive material, but I wonder if anyone does this for the small servo pots  used in our systems.

        IMHO, the processing speed is not as much of a problem for us, at least  in pattern, as we might think it is. It is much more important that the  surfaces end up at the same position, since that will affect the flight more  than the small difference in position *while the surfaces are moving*.

        Bob R.

  brian young  <brian_w_young at yahoo.com> wrote:

            I think this is a problem with the processing capability of the TX. You  have a progammed mix, it reads one channel then the next. I had a 9CAP where  this problem was very noticable. Now on my 9Z its not as noticable, it looks  like the slave servo is a little slow to catch up. Only real solution in my  opinion is a match box, y-cable or similar device.

        I can sit back and pick which servo is the slaved just watching the  surfaces. 

        Im understanding that your endpoints match when the surface are  finished moving, and the middle is matched as well. You should be able to stop  at any point in the travel and the pointers line up, when they stop moving. If  this is the case your linkage would be ok. 

        I think the speed of the servos has passed the processing speed of the  radio.

  Bob Richards  <bob at toprudder.com> wrote:

            That might be the cause, or it could simply be a difference in  linearity between the servos.

        Bob R.

    _______________________________________________
NSRCA-discussion mailing list
NSRCA-discussion at lists.nsrca.org
http://lists.nsrca.org/mailman/listinfo/nsrca-discussion
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.nsrca.org/pipermail/nsrca-discussion/attachments/20060812/f6d93fe8/attachment-0001.html 