Conference vmszoo::rc

    
> First, could someone please explain a servo in it's intricate little
> details?  I've seen somewhere in this file that you feed them power
> and a 'control' signal, but what is the control signal and what
> is the mechanical response?
    
    In all of the servos I've used, you give them +5v, ground, and a
    TTL level control signal.  The control signal has an active-high
    pulse once every 20 to 25 milliseconds.  The width of this pulse
    determines the servo position.  In my Futaba, the pulses are roughly
    0.75 microseconds wide for one end of the servo throw and 1.75
    microseconds wide for the other end.  WARNING : each servo manufacturer
    has a different pinout for the 3 wires that go to the servo!  It
    is easy to apply reverse voltage if you're not careful.

    Inside the servo there is a DC motor that turns the output shaft
    and a pot linked to the output shaft.  The resistance of the pot
    is the feedback signal to the servo loop.  The controller generates
    an analog voltage proportional to the width of the incoming pulses,
    compares that to the feedback and drives the motor accordingly.

    With standard servos you get about 90 degrees of total throw, with
    retract servos you get more like 180 degrees I think.

> Also, I imagine most of them have a
> rotary output, with a mechanical linkage to turn that into a push
> or pull, is this right?

    Right.
    
> Also, I'd really appreciate a technical discussion of radio controls
> and what the control signals that are passed back and forth
> are.

    For the non-PCM systems it's very simple.  In the transmitter, each
    channel has a pot whose resistance is proportional to the stick
    position.  The encoder creates a pulse for each channel, the width
    of the pulse being proportional to stick position.  The pulses for
    each channel are time-multiplexed to a single pulse stream, and
    this repeats every 20 milliseconds or so.  I.E. - every 20 milliseconds,
    there are seven pulses (for a seven channel system).  The only other
    thing the transmitter does is modulate a carrier (AM or FM) with
    this pulse stream and transmit it.  With the AM systems they use
    100% amplitude modulation, i.e. when the pulse is "high", the
    carrier amplitude is zero.  I don't know what modulation frequencies
    they use in FM.
    
    The receiver demodulates the pulse stream from the carrier, then
    demultiplexes it into one pulse stream per channel  (a one-shot
    is used to sync-up with the beginning of each new pulse stream
    and start the demultiplexer at the right place).  Each signal is
    then sent to the appropriate servo, which will convert pulse width
    to position (a linear relationship). The end result is servo arm
    position proportional to transmitter stick position.
    
    I can't tell you specifics about how they do it with PCM, as I
    haven't opened one up.  I would assume they have an A/D converter
    in the transmitter which samples each stick position, producing
    an n-bit binary value for each channel.  This would then be time-
    multiplexed to combine all channels in a single bit stream and
    modulated on the carrier.  The receiver would then demodulate the
    carrier, sync on the signal, and D/A convert each n-bit value
    for each channel.  I assume the resulting control signal to the
    servos is the same as the old way, but can't swear to it.
    

> 1 - Manual operator controls read by control point computer:
>        
> 2 - Base computer talks to remote computer via radio link,
>           maybe along something analogous to an RS-232 link (as
>           I know how to handle that):
>           
> 3 - Remote computer controls servos and such.
>            
>     Can this be done, and if so, what's the highest baud rate (or
>     equivalent) I could get out of the link?
           

    I'm not sure you need the remote computer.  Just have the base
    transmit the servo position information like RC transmitters do.
    Then all you need remotely is the receiver and servos.  You *could*
    do a remote computer link using this kind of system at slightly
    better than 300 baud, it just depends on how much effort you're
    willing to put in.  One word of warning - you must use ground
    frequencies, not aircraft frequencies, and if you do up your
    own computer link you must make sure the newly modulated signal
    does not violate adjacent channels.  Hope you have access to
    a 100MHz spectrum analyzer!

    Good luck, hope this helps.
    
    Bill

    P.S. look at the Signetics NE5044/5045 encoder/decoder chip set
    	 for an idea of how the encoding and decoding schemes are
    	 implemented.  These chips are used in my Airtonics system
    	 and are in my Signetics data book.

    Thanks, Bill, that was just what I was looking for!  I do have access
    to a spectrum analyser, but I'm not sure I want to play those kind
    of games.
    
    I always thought that each 'channel' used a different radio frequency,
    but it sounds like everything is just time multiplexed onto a single
    bit-stream.  How fast is that single digital channel allowed to change
    state?  Would it be OK to use your own coding?  I sort of doubt you
    could stuff a bit-stream from a UART into it without some kind of DC
    restoration (Manchester or something), so I guess maybe synthesizing
    the pulses at the base end would be better.  How many channels can you
    get in a single radio?
    
    Thanks, I'll look at those chips and see if this makes any sense!
    
    Willie
     
        

>    I always thought that each 'channel' used a different radio frequency,
>    but it sounds like everything is just time multiplexed onto a single
>    bit-stream.

    	Right!  That's why they only need one carrier per system.
        
>    How fast is that single digital channel allowed to change
>    state?  

	If you want to stay within one channel you have 40Khz
    	bandwidth to play with until 1991, when it becomes 20Khz.
    	Of course you could use more channels if you're not going
    	to be sharing air-space with other RC'ers.  Come to think
    	of it, I'm speaking for the air frequencies.  I don't know
    	how the car/boat frequencies are spaced, or how 1991 affects
    	those frequencies.  Any car/boaters out there know?
    	
> Would it be OK to use your own coding?
    
    	Sure, the RC manufacturers don't own the encoding scheme.  All
    	you have to do is stay within the bandwith limits.
    	
> I sort of doubt you
> could stuff a bit-stream from a UART into it without some kind of DC
> restoration (Manchester or something), so I guess maybe synthesizing
> the pulses at the base end would be better.  How many channels can you
> get in a single radio?

    	The most I've seen is 8 channels, usually for helicopters.
    
> Thanks, I'll look at those chips and see if this makes any sense!
    
	No problem - if you need a copy of the specs let me know, I'll
    	send them.
    
    Bill
        

    There is an error in my first reply (.1) ...
    
>       determines the servo position.  In my Futaba, the pulses are roughly
>    0.75 microseconds wide for one end of the servo throw and 1.75
>    microseconds wide for the other end.  WARNING : each servo manufacturer

	"Microseconds" should read "Milliseconds".  Oh well, I was only
    	a factor of 1000 off!
    
    Bill

    Thanks Bill, I found the chips and looked things over, and between
    the ancoder, the decoder, and the servo chip, it all falls into
    place.  Looks like my best bet is D/A converters driving the
    transmitter box unmodified.
    
    Willie

    Bill,
    	
    I thought those pulses were narrow.  You could get lots of channels
    in there with a 20ms update rate and less than 2 us per channel.
    Somehow, I think you might need more than 20Khz bandwidth in the
    channels though.
    	What's a milli-micro between friends
    

    Turns out the pulses are narrow, but the Pulse Width Modulation
    is done on the space between pulses, not the width of the pulse.
    The timing in the Signetics book looks a bit bizarre till you figure
    that out, but then it kind of makes sense.  At 20 KHz bandwidth,
    you should be able to move a 10 kilobit ssignal (or 9600 baud, if
    you want a standard baud rate), but I doubt it would be that easy.
    
    Willie

    The modulation is not done in the spaces.  The width of the pulse
    is the position information for that channel.  The chip in the servo
    converts the pulse width to a position command for the servo control
    loop.  The pot in the servo is the position reference for the output
    shaft.  In the standard servos, the pot is hooked directly to the
    output shaft, and the motor drives the output shaft through a large
    gear reduction.  I would stick to the standard modulation scheme
    so that you don't have to fool with the servos.  You could send
    additional information over the channel in the time when the pulses
    would be sent for unused channels.  If you maintain the required
    spacing between frames that is required to sync up the receiver's
    demultiplexer, you could send anything you want for most of the
    frame if you need only a couple of RC channels for servo control.
    
    CHarlie
    

    Well, we are both about half right.  Looking at the Signetics spec
    shows that the encoder generates a signal that encodes the pot position
    (or input voltage, which is usually the same thing) in the spaces
    between the pulses, but the decoder sends a signal to the servo
    in which the pulse width does the encoding.  As long as it all works,
    the exact details of the 'magic' in between aren't too important,
    input voltage or pot position translate directly into servo position.
    I'll probably end up sticking with the standard modulation scheme.
    
    Willie

    One more radio question:  What are 'dual rate' radios and why would
    you want one?
    
    Willie
    

        Re:< Note 394.11 by SNDCSL::SMITH "William P.N. (WOOKIE::) Smith" >

                Dual rate  transmitters  allow  you to select between two
        different  levels of  control  surface  throw  at  maximum  stick
        deflection,  usually  on  just two controls.  I think  it's  most
        normal  on  elevator  and ailerons.  It's extremely useful to  be
        able  to  select  between very high sensitivity to stick movement
        and a lot less.  I usually put it on low when I instruct and keep
        it on high  for  myself.  The most common mistake I have observed
        is  giving  too  much  control,  particularly  when  in  trouble.
        
        Anker

Its kind of difficult to see from the diagrams, but the actual
transmitted pulse is about 300 microseconds wide.  The spacing
between the LEADING edges carries the position information, and
varies in theory from 1 to 2 milliseconds, with 1.5 milliseconds
being neutral.  The reason that there are only 8 channels in the
20 or so millisecond (it can be as little as 15 milliseconds)
frame rate is that there is a 10 or so millisecond pause between
frames.  So the transmitter puts out the eight 300 millisecond
pulses then "rests" for a 10 millisecond period.  Of course, this
is the modulation; the carrier is on for the full period.

The pause is called the sync pause and is used by the decoder to
determine where channel 1 starts.  The decoder contains a missing
pulse detector to find the pulse and reset the channel decoder
string.

BTW-- as I found out the hard way, if you only transmit, say, 3
channels, the 15-20 millisecond frame rate still holds.  This is
because the servo decoders are set up for that frame rate.  They
go crazy if you transmit information pulses to them at a higher
rate. 

Fred Marks' book Getting The Most From Your Radio Control System
is invaluable in its coverage of technical detail.  And its
straight from the horse's mouth. 

    
>    Its kind of difficult to see from the diagrams, but the actual
> transmitted pulse is about 300 microseconds wide.  The spacing
> between the LEADING edges carries the position information, and
> varies in theory from 1 to 2 milliseconds, with 1.5 milliseconds
> being neutral.  The reason that there are only 8 channels in the
> 20 or so millisecond (it can be as little as 15 milliseconds)
> frame rate is that there is a 10 or so millisecond pause between
> frames.  So the transmitter puts out the eight 300 millisecond
> pulses then "rests" for a 10 millisecond period.  Of course, this
> is the modulation; the carrier is on for the full period.

    Right you are!  I guess my memory didn't serve me right - I
    just took another look at my Futaba on a scope and the modulated
    waveform agrees with what you said.  Oh, and thanks for saying
    "300 millisecond" when you meant "300 microsecond", I feel
    better now!
    
    Thanks for the correction -
    
    Bill
    

    re: previous discussions about transmitter modifications
    
    Remember that radio transmitters are regulated by the FCC (Federal
    Communications Commission). There are regulations about frequency,
    power, and encoding/encryption of information. I am not familiar
    with the specific regulations around the 72/75 MHz R/C channels,
    but my recommendation is that if you don't know the regulations
    you had better not modify any transmitter. 
    
    Dave Hughes

    How would you define 'modifying the transmitter'?  If I have a computer
    driving a D/A converter, connect the resulting voltage to the encoder
    where the pot goes, so I can 'fly by wire', would that be legal,
    or am I supposed to stay completely outside the case?
    
    Willie
    

    Also, if I buy a 7 channel radio, do I get 7 analog channels, or
    do I get 4 analog channels on 2 sticks and 3 on/off channels?
    
    Is it possible to buy the transmitter and receiver as a package
    without the servos?  I want to get the best, highest power, ball
    bearing servos (maybe sail winch servos or something), which would
    mean I would have to toss the servos included with the package...
    
    Willie

    re: .16
    
    I would be careful about changing anything that drives the transmitter
    section. I just don't know what the regulations are, so I can't
    advise what would be ok. Changing the input to the modulator could
    cause problems in the transmitter portion. For example, rise times
    of input signals alter the bandwidth of the transmitted signals
    - there are specs for the rise and fall times. Unless you're sure
    you know what you're doing, I'd stay out of the case.
    
    Regards,
    Dave

    I'm not sure I'd be modifying the transmitter section or anything
    that directly drives it.  I wan to change from:
    
    	A pot that delivers a varying voltage level to the encoder (that
    in turn drives the transmitter).
    
    				to:
    
    	A D/A converter that delivers a variable voltage to the same
    place that the pot did.  The only difference is that the rise times
    of the signals into the encoder (not the transmitter) could probably
    be faster.
    
    	When I speak of the encoder I mean the thing that takes analog
    signals from joysticks and pots and produces that time multiplexed
    PPM/PWM signal that is then sent to the transmitter.  I suppose
    I could always bring it in and check with our new spectrum analyzer.
    
    Willie

.20: The editor is illiterate; a common fault...

.21: I've used the ACE transmitter module several times.  When
you buy it you will get an instruction sheet that tells you the
modulation voltage level (I think its 7V; its been a while).
As you probably know, exceeding the specified modulation level is
very much a no-no with the FCC.  As long as you don't do this,
and don't fool with the transmitter tuning coils, you are in Fat
City.

Is that module good???  Well, ACE claims that don't need to
change it to meet 1991 specs!  Its the one I'll put into my
Futaba when the time comes.

Also worth considering is their "digital commander" modulation
board. I think they're selling it as an experimenter's board.
This is a simple way to get the AD's output into the transmitter
in an acceptable form.  When I built my transmitter I used this
circuit but I modified it for CMOS chips.  It is very resilient
to being fooled with and produces a nice output waveform.  

>  A D/A converter that delivers a variable voltage to the same
>  place that the pot did.  The only difference is that the rise
>  times of the signals into the encoder (not the transmitter) could
>  probably be faster.

	I don't think you'll have any problem with this.  The encoder
    sees the pot output, or your D/A output, as a DC level (relative
    to the frequency the decoder is running at).
    
    RE: the question about 7 channel systems, most of them have 6 analog
    channels and one on/off (retracts).  Four of the channels are on
    joysticks and two on slide pots (on top, side or front of xmitter).
    

    Bill

    While we are on the subject of servos, what is the usual servo traval
    in degrees?  I've seen some servos specified in time to make a 60
    degree swing, but there must be more traval available.  I'm
    specifically interested in the servos that come with (or are an
    extra price option, precision servos?  What's the difference?) the
    Futaba 7 channel FM dual-rate radios, as I'm planning on ordering
    one in the near future.
    
    Willie

    I have noticed that there are several radio "kits" on the market
    here in the UK and I'd guess the same is true just about everywhere.
    Question: how difficult are these things to build and what do you
    get for your money (ie. do they make economic sense)?
    
    Keith
    
    P.S. I joined DEC on the strength of my S/W background although
    I am not unfamiliar with a soldering iron...

Keith,

The average kit is rather straightforward and reasonably easy to assemble,
assuming some familiarity with electronic component identification, average 
dexterity and [**most importantly**] above average soldering skills and know-
ledge regarding handling of delicate solid-state/IC components.  As a rule, the
critical RF section(s) is/are preassembled negating the need for a scope and 
other exotic [expensive] test equipment for final tuning/alignment though some 
basic equipment, e.g. VOM/DVM, will be required.

I've built many electronic kits from my high-school days on and consider myself
[in all humility] very good at it and a more than adequate solderer, having 
taught mil-spec soldering on several occasions. And, I've built a coupls servo's
and various [ACE] pieces of support equipment [tachs, chargers, etc.] but I much
prefer leaving the assembly of my radio to the factory who have automated assem
bly/soldering equipment and the QC and testing resources [unavailable to the kit
builder] to assure a quality system.

The cost differential is too insignificant for me to consider; a complete, 
ready-to-fly, kit built system will cost within $20-to-$50 of a factory-built 
item and "can," believe it or not, frequently cost "more!"  Look at the price of
the new generation of "Challengers, Conquests," etc.; you just can't build a kit
that cheaply.  It's just my opinion, but I prefer to spend the extra coupla' 
bucks and apply my expertise to airframe construction and the homework, i.e. 
maintenance, care and feeding of the system, etc., required for crash-avoidance.
I've only lost 4-aircraft in nearly 25 years of radio modeling [no boast, just 
fact], which, I believe, speaks well of factory built radios so I see no reason 
to "do it myself."

Adios cuzzin',	Al

Having built the ACE Silver Seven receiver and a couple of ACE
Bantam servos, I have to agree with Al; the time could be better
spent.  Especially with those nice Airtronics servos on the
market for less than the ACE servo kit!  The only thing I will
add to Al's commentary is that the work is small and delicate.
The receiver board is packed solid with components, and the servo
board is just as bad.  I use an industrial grade soldering iron;
if I had to use the garden variety soldering iron it would've
been much harder.

My eyes have dictated that I've built my last servo board, but I
may tackle the ACE 1991 receiver in a couple of years.  

    One more silly question about radios and then I'll buy one and find
    these things out by playing:
    
    	What frequencies would you reccommend for ground use in the
    Acton area?  I'm not planning on racing against anyone, but would
    like to stay away from busy frequencies or those with lots of
    interference from other sources.  Should I go for one of the channels
    that just opened up, or what?
    
    Willie

    Willie,
    
    First ** AND "MOST" IMPORTANTLY ** , you !! MUST !! buy a radio
    on a frequency dedicated to surface [boats/cars] use ONLY!  If you
    already knew this, excuse the reminder but this is EXTREMELY im-
    portant!  Yer' best course for determinig best/least used/clearest
    frequency is to ask around the local tracks, hobby shops, etc. or
    simply observe which seem to be the busiest channels, most trouble-
    free, etc.  I realize that that's what yer' doing in the notes_file
    but there's no substitute for observing what goes on in the actual
    dynamic environment to intend to operate in.                      
    
    All [legal] channels currently in use should be exclusive-use
    frequencies but some "may" be better than others...the only way
    to be sure is to inquire of the modelers who are currently familiar
    with yer' particular local situation.
    
    Adios,	Al

    Al,
    
    	Thanks, I was aware of the 'surface only' requirement, I don't
    want to be dropping planes out of the sky that I'm not even aware
    of.  I'm not really into the modelling community, so I don't know
    what hobby shops are in the Acton area, and I've only talked to
    a couple of other modellers, neither of whom are very near here.
    I'm initially planning on running the model around the grounds here
    at my apartment complex and maybe graduating to the sand pits out
    back, so I probably don't have to worry about conflicting with other
    surface vehicles.  I was hoping some of the Acton residents who
    read this file might have some hints on frequencies.  Worst case
    I'll just call Tower and see what they have in stock, starting with
    the higher channels (the ones just recently allowed).
    
    Willie
    

That brings up something I've been avoiding, the subject of servo
maintenance.  The most likely reason the servo locks up is that
the feedback loop from the pot is broken due to a dirty pot, or a
defective one.  A secondary reason is that the transmitter is
putting out an out-of-tolerance positioning signal, hardly
likely in this case -- but I've bent a few servos that way in my
experiments. 

The "books" say that you should clean and lube the pot yearly;
I've had a guilty conscience for not doing this -- now I strongly
suspect that no one else does it either.  But if there is a
true-blue soul out there, who does lube pots, (and flosses
regularly): What do you use for lube, and just how much trouble is
it?  I know that you can buy the lube from ACE -- is that the
only place to get it?

It may be a couple of days before I see this reply, I've got to
see my Perodontist.  If you don't know what that is, you're not
as old as Al. 

    John,
    
    Back in the dark ages when we were running Bonner Transmite/Digimite
    servos, maintenance was a "verrrry" necessary evil.  At least once
    a month [during active flying periods], you had to take all the
    servos out of the plane, disassemble them and clean the wire-wound
    feedback pots with alcohol, then lube with a fine film of silicone
    grease and reassemble and reinstall in the plane.  This was a nasty,
    time consuming "CHORE" but it was mandatory if you wished to prevent
    the guaranteed crash which "would" occer if you got lazy.
    
    With the advent of ceramic pots, carbon button wipers, etc., etc.,
    I, quite frankly, thought servo maintenance/pot cleaning was a thing
    of the past and, indeed, I haven't cleaned a servo in years,
    subscribing to Kevin's "If it ain't broke, don't fix it" theory.
    Now, either I've been extremely lucky or the "books" are exaggerating
    the need for such maintenance.  I, quite honestly, can't remember
    the last time I had a servo fail, for "any" reason, and would have
    to be convinced of a "real" need before I'd consider returning to
    the distasteful chore of servo maintenance.
    
    I fly Futaba, exclusively, and, perhaps, other makes of radios/servos
    are less tolerant than Futaba, but I can't admit to knowing "anyone,"
    including the "ultra"-cautious Kent Walters, who does any kind of
    routine maintenance on their servos.  I remain to be convinced of
    this necessity.
    
    Adios,	Al

    	The pots on modern servos are not designed to need any PM. 
    If a pot gets flakey causing the servo to chatter or hunt, it should
    be replaced.  Vibration is the major cause of servo failures due
    to gear wear and possible broken wires.
    
    Charlie
    

    Just thought I'd add a quick note here on the Futaba transmitter.
    This has to be the _most_bizarre_ PC board it has ever been my
    priviledge to encounter.  When you open the case you are confronted
    with a cheap looking single-sided phenolic PC board with no
    plated-thru-holes, no solder mask, no silk screen for component
    locations, just normal parts and lotsa jumpers (to make up for the
    fact that it's only single sided).  Carefully working around the
    rats' nest of wires and pulling the board out gives you a view of
    the back side of the board, which transports you to a completely
    different world:  There you find surface mount parts, solder mask,
    surface mount jumpers_for_gloz_sake, and one of the highest-tech
    looking pieces of hardware around.  Really wierd....
    
    BTW:  The transmitter module wasn't too hard to figure out with
    a scope, does anyone have a source for the NE5044/NE5045 chips so
    I can take a shot at building my own encoder module?  It's looking
    like that might be a bit easier than trying to interface to the
    Futaba encoder....
    
    Willie

Both Ace and Digi-Key stock them.  Also get a copy of Fred Marks'
book Getting the Most From Your Radio Control System.  He has
extensive information on the Silver Seven encoder/decoder which
he designed.  I assume you have the Signetics info, if not, let
me know.

Last night I was installing the aileron servo on a PT-40. The instructions call
for mounting the pushrods at +/-45� off center so that you get a "differential"
motion control on the ailerons, ie., the ailerons will travel further in the up
direction than in the down direction. The problem is, in this configuration, the
maximum servo rotation one would want is +/-45� while the servo actually wants
to deliver +/-90�. If you picture the pushrods mounted on the wheel at 7:30 and
at 10:30 (on a clock face) and the trailing edge to the right, at 3:00, if I
rotate clockwise more than 45� the one at 7:30 goes past 9:00 and actually
starts to go in the opposite direction again. At a full 90� rotation, I end up
with full up on one of the ailerons and neutral control on the other one. Is
this correct, or should I do something to limit the travel of the servo? I
thought of modifying the resistance of either the pulse width forming circuitry
in the transmitter or of the feedback circuitry of the servo. What I need (I
think) is a dual rate capability for the transmitter, but that isn't one of the
features of the Futaba 4 channel Conquest. 

Dave

    Your servo goes +/- 90 degrees? The ones I've seen only go +/- 45.
    Have you actually verified that your servo goes +/- 90?
    
    						John.

    Most servos will turn much further by hand than they will when driven
    by a standard transmitter.  If you do have a servo that will do
    180 degrees when driven 'normally' I'll trade you even for an S-148!
    
    Willie

    This question may apply more to car RC than airplane because of
    the reverse switches I have seen on airplane RF transmitters
    (maybe there the same switches on the CAR RC pistol controllers
    and I havent seen them).
    
    If using the servo signal to control a motor speed controller,
    which way would be forward and which way reverse? 
    
    I am assuming for the moment that 2.0 msec would be max forward,
    1.5 neutral and 1.0 max reverse.   The only reason this matters
    is that I plan to provide 4 functions for motor speed control
    with this one signal:  forward linear from about 1.6-2.0msec, neutral 
    binary from about 1.4-1.6msec, brake binary from about 1.2-1.4, and 
    reverse analog from 1.0 to about 1.2msec.    As you can see simply 
    reversing the motor leads would not make things right if I guessed wrong.

    I am definately at a disadvantage designing this thing without 
    a car, transmitter and receiver in hand. 

    Thanks,
    Walt

    Walt,
    	Most radios have reversing switches on the transmitter that
    make it possible to have throttle movement affect the pulse width
    in either direction.  I would guess that the most straight forward
    way would be increasing pulse width = increasing forward speed.
    You want to put a delay on reverse to prevent damage to the car
    if the stick is shoved back.  The futaba speed control has maybe
    one second delay so that you get instant braking but not reverse.
    
    Charlie
    

    
    Eric stopped by to play yesterday at lunchtime.  We happily noted
    that it was cold and rainy outside, so we wouldn't be flying even
    if the hunters weren't stalking the CMRCM field!
    
    He brought along his JR PCM-9 with the intent of adding a slow-down
    circuit like I had put in my Futaba.  Charlie tore himself away from
    work to join the fun, and we ended up drawing a pretty good crowd
    of people who are sick of debugging disk drives.
    
    We ripped the PCM-9 apart and quickly found that the retract channel is
    a digital input (as opposed to my FGK, which is an analog input).
    We concluded that we would have to take one of the analog aux channels
    and wire it to a retract switch so we could add the RC delay circuit.
    
    At this point Eric decided to take a closer look at the ACE servo
    pacer he had purchased, and was surprised to find that it performs
    the exact functions he needed.  We left the PCM-9 in its original
    configuration.
    
    The servo pacer is used with a regular servo, not a retract servo.
    I learned yesterday that retract servos are not proportional, just
    one end or the other.  So it's impossible to slow them down without
    modifying the servo itself.  So how do you get the additional throw
    of a retract servo (close to 180�) with a standard servo (about 100�)
    so you can slow down the rate?  Happily the answer was right under
    our noses.
    
    The servo pacer detects from the incoming pulse stream when you have
    thrown the retract switch,  and synthesizes _new_ pulses to the
    servos.  So the servo end-points are not limited by the pulses
    sent out by the transmitter, they are set by two pots in the servo
    pacer.  These pots allow close to 180� throw.
    
    If your servo does not have mechanical stops that prevent it from
    going more than 100�, you are all set.  Using the pacer, the standard
    JR servo was able to travel about the same amount as the retract servo.
    
    A third pot in the pacer sets the rate of movement, so you can make
    the retract last several seconds if you want.
    
    The servo pacer has the disadvantage that you have to add something
    inside of your airplane, so I still prefer the transmitter solution,
    but it is a neat little device, small and lightweight, and not much
    circuitry.
    
    Bill

    Does anyone know of any device(s) on the market that allow you to
    control a servo without the tx/rx turned on??
    
    I had all kinds of engine problems at Rhinebeck and it would have been
    great to work on the engine without tying up the frequency...
    
    (read-only noters, feel free to send me mail on WRASSE::FRIEDRICHS)
    
    thanks,
    jeff
    

    Hi Jeff,
    Ace RC used to make a box that you could plug a servo into and control
    it without needing a transmitter and reciever. I can't find my ACE
    catalog or I'd tell you whether or not they still make it and how much
    it cost.

>>    Does anyone know of any device(s) on the market that allow you to
>>    control a servo without the tx/rx turned on??
    
    There are two classes of such devices --- one is a servo exerciser, and
    the other is a local manual control.  The  technology is trivial, and
    there are devices on the market.  e.g. 

        Automatic Servo Exerciser (kit)	$18 to $35  (what the market will bear)
        H&N Electronics			619-373-8033
        10937 Rome Beauty Drive
        California City, CA  93505      

    I would guess that to scratch build a manual servo control would cost 
    about $4 plus [old] battery pack plus connectors. It would have a dual
    555 to generate a pulse every 20 ms and to vary (via a potentiometer)
    this pulse in width from 1 to 2 ms.  This is what you want.

    I've considered building a random exerciser to test servos for dead
    spots after a severe landing.  This would generate a random pulse
    width, hold it for a second, and go on to another random width.  In an
    hour it should test at least once almost every possibility for a dead
    spot.  

    The two functions can be combined in one device.  I'm not sure exactly
    what the device from H&N Electronics does.

    Alton, who has more projects on file than time to do tiddly

    My Airtonics Chanpionship raidios both have this feature built in. 
    There's a plug on the back of the TX that you use with a cord that's
    included with the radio.  You plug the cord into the charge plug and
    leave your tx turned off.  Now, when the receiver is turned on, you
    have control of all of your servos without transmitting any RF.  THe
    only problem with this is most clubs require that your TX be in the
    impound unless you have the frequency pin.  I've never used the feature
    for this reason.
    
    Charlie
    

    Here's what I've found, doing the autopsy on the four Airtronics
    94141 servos that were in my Legend crash:
    
    The two all-metal gear,later model came through okay. Actually,
    even these use a plastic gear as the first one after the motor out
    put shaft.
    
    The earlier model used a second plastic gear, further downstream
    in the middle of the gear train. This second gear stripped in two
    servos.
    I bought two replacement gear sets (STICKER SHOCK ALERT !) @ $20
    per, half the price of a new servo.
    I thought I could replace only the stripped plastic gear with the
    new metal one, and save the other gears in the set for later use.
    BUT, I discover that the base of the final output shaft, the one
    with the splines on top for the arm, was also bent causing slight
    but noticeable wobble.
    So I replaced all the gears in both servos.
    Now arrives a baffling problem that spent 2+ hours diagnosing:
    
    After reassembling with the new gears, both servos would rotate
    smoothly, but only in one direction. When they reached full travel
    one way, they would lockup solidly. I disassembled/reassembled
    many times, compared with a known good servo, etc. and just
    couldn't see anything wrong.
    In desparation I plugged them into the rcvr. and powered up.
    Now, one works perfectly, the other is stilled jammed. The good
    one under power, still jams under hand rotation. What the....?
    
    Anyway, it turns out that the recessed hole in the top of the case
    that receives the top of the gear shaft in the next-to-last output
    stage, gets elongated slightly with sufficient side load impacts.
    The gear train itself will be working ok, but the shaft cocks
    over in the elongated hole and binds up the gear train when it reaches
    full travel on one side.
    If the elongation isn't too bad, it will work under power but not
    by hand twisting.
    The fix is to replace the top of the case. Cases are ~$3.50, so
    it's off to the hobby shop to order two new ones.
    
    Any crash hard enough to strip gears or break output arms , warrants
    close investigation in this area, even if no other symptoms arise.
    
    Terry

    I decided to use servo savers on my flap servos on the new Legend.
    
    I have the CRP, or CPR, or PRC brand, whatever, they come in a yellow
    package at $3.00 per and fit all Airtronics non-micro servos.
    
    The body of the servo saver where the screw goes through is too
    thick to use the stock screw that comes with the servo.
    The 94141 uses a machine thread *3 MILLIMETER* screw because of
    the metal output shaft.
    At my local hobby shop I was able to find a 3 mm machine thread
    screw in the R/C car section where they have boxes of them; servo
    savers being all the rage on R/C cars. They are twice as long as
    the stock screw and go down into the output shaft a safe distance.
    
    Terry

    Somewhere in here is a discussion of the metal gear Futaba micro servo
    announced last year. They've recently come available for about $53
    (list is $79.95) I got a futaba catalog at the WRAMS show and they use
    the same case as the s-133 so the answer is yes, you can upgrade your
    133 to metal gears. (same replacement part number for the cases). The
    bad news is that the catalog lists the metal gears at $39.95 list!

    Another data point on the 5102 metal gear micros. Tower has them listed
    for $48 in the latest Tower Talk. Still no "actual" price on the gear
    trains themselves for retrofitting 133s.

Last night I tore down my stripped S9601 servos to replace the stripped gears. 
These are coreless servos with one metal gear. In all three I stripped it was 
the main shaft gear that was stripped. This gear had two holes 180 degrees 
apart that tabs from the motor fit into. 90 degrees from this is a "bump" that 
sticks up. The place where they all stripped was opposite the bump. the bump 
prevents you from rotating the gear 180 degrees to fix the problem. Since the 
servo only moves this main output gear under 180 degrees, I trimmed the "bump" 
off carefully with a knife and reinstalled it 180 degrees from original. The 
missing teeth now don't come in contact with the rest of the gear train and I 
get a second life out of the output gear. You can only do it once but since I 
had three stripped servos and two replacement gear sets, I figured it was worth 
a shot. Moving the servo with the case top off showed that the servo didn't get 
to that point. This bump provides the mechanical stop when the servo is 
unpowered. This is also a trick I hope to remember when I strip one at a 
contest since it only took 10 minutes to do and could be done as a field repair.

        Re:     <<< Note 394.55 by GAUSS::REITH "Jim 3D::Reith MLO1-2/c37 223-2021" >>>

        Jim,
        
                I did  exactly  the  same thing at a contest 2 weeks ago!
        It works like a charm.
        
        Anker

Yeah, the only possible problem I can see is that it allows over rotation 
of the servo when unpowered which might be a problem. It wouldn't be a 
problem if the surface had some mechanical stop that prevented it but as 
a loose servo you need to watch it.

    As good a note as any...
    
    I've got a few extra s148 servos that I would like to convert to sail
    winches. I can get the behavior I want (multiple revolutions both
    clockwise and counterclockwise, neutral stick = stopped) by cutting the
    feedback pot shaft and cutting the rotation limiter stops off the final
    gear. My only issue now is that the 'deadband' is still pretty small. 
    I would like to increase it so that only major stick movements would
    cause the servo to move. Ideally, the trim on the transmitter should
    have no effect - the servo would remain motionless regardless of trim
    setting. I figure that the deadband can probably be changed by new
    resister and/or capacitors on the board, but I have no schematics and
    I'm left with 'reverse engineering' (read: trial and error) to get the
    desired deadband.
    
    Any constructive advice?
    
    Loren
    

hi loren,

if you look at the IC on the pcb  and see it of type NE 544 i can supply
schematics/info which is the deadband r/c combination.

else only try and error will help, but with the few components you should
be able to do it that way as well.

cheers
joe t.

    We used a 10 turn pot instead of the original. This is hooked to the
    end of the wintch pullies, servo on one end, pot on the other with an
    extention wire feeding back into the servo electronics. 
    
    You take out any stop teeth on the gears in the servo. This will give 
    10 turns of your servo for the same movement on your stick making it 
    less sensitive. We did this type of system when we first started sailing 
    in the early 70's using a kraft set, I believe a 15H servo, Maybe 
    wrong it was a long time ago.
    
    I would think that the 148 would be pretty marginal power,although a
    pully will require less power than an arm. 
    
    We sail Marble Heads (52 inches) and for sure a 148 wouldn't have 
    enough power.
    
    Happy Sailing
    Fred
                                        
    

    The 10 turn pot seems just the ticket for my application. Great idea.
    Thanks...