[NSRCA-discussion] SD-10

John Pavlick jpavlick at idseng.com
Tue Aug 25 06:58:03 AKDT 2009


"The point I want to express is analogous to making your house more secure by putting 4 giant deadbolt locks...on a hollow core door.  You're only as safe as your weakest point of entry. Be it kicking through the door or breaking a window...the locks are more than adequate and no longer the weak point in the system.  I believe the same is true for resolution, it's no longer the weak point in the system."
 
 Good point. FYI - I only have one lock on the  door, but I have 6 bullets in the gun! LOL
 
John Pavlick


--- On Tue, 8/25/09, Atwood, Mark <atwoodm at paragon-inc.com> wrote:


From: Atwood, Mark <atwoodm at paragon-inc.com>
Subject: Re: [NSRCA-discussion] SD-10
To: "General pattern discussion" <nsrca-discussion at lists.nsrca.org>
Date: Tuesday, August 25, 2009, 9:59 AM


Hey Tony,

I don't disagree with your math or understanding, though I'm trying to make a slightly different point I think. I also misspoke in a previous post regarding how the resolution and EPA settings interact, you're explanation is clearer and is also my understanding.

I also want to state that I'm not advocating that 2048 isn't "better".  The original question was "Why didn't ATX choose 2048 over 1024" and the answer from their end was cost and the "opinion" that given the cost choice between speed and resolution, that speed had a more noticeable impact on most setups.

The point I want to express is analogous to making your house more secure by putting 4 giant deadbolt locks...on a hollow core door.  You're only as safe as your weakest point of entry. Be it kicking through the door or breaking a window...the locks are more than adequate and no longer the weak point in the system.  I believe the same is true for resolution, it's no longer the weak point in the system.

To follow your math/understanding, the "step" size for a given resolution is constant.  So in your max travel example, which I think is a good one, we're probably talking about spreading 1024/2048 across 135 degrees of throw at max.  Thus the "step" size for 2048 is approx 1/15 of a degree, and therefore at 1024 it's approx 1/7th of a degree.  Regardless of how you dial down your rates, that will always be the smallest step size.  I.e. if you reduce rates as you suggested, reducing EPA to 100% from 150% (down to 90degrees) and D/R down to 80% from 100% (down to 72degrees of rotation) we now have 1080 pts across 72degrees with 2048, or 540 pts across 72degrees with 1024.  In Both cases the step size is still 1/15 of a degree or 1/7th respectively.

My assertion is simply that 1/7th of a degree of resolution (approx .0012" movement on a 1" servo wheel) is smaller than other area's of inaccuracy in our setups.  Or at least in most people's setups.  Getting a setup tighter than that is hard.  Keeping it there is even harder. Servo gear trains have slop, ball links, clevises (big slop), wing adjusters, even the wing tube all allow for movement in micro amounts that ruin our accuracy at a given point. And I won't even mention the variance in deflection if you're flying glow due to vibration which I'm sure makes this whole conversation moot.  (I know I know...I need to convert to Electric...you keep telling me! I'm working on it!).  

Anyhow, I hope you understand my perspective. People are choosing where best to spend their money to gain the most functionality and advantage in their flying setups, and in a similar fashion Airtronics did the same.  They felt they were giving the pilot the best bang for the buck by choosing speed over resolution.  I DO think when they complete their flagship radio (a ways out) it will likely be 2048 so as to be comparable with the other top of the line systems.  I'm also pretty sure it won't be $500 though.

This has been a great discussion though and hopefully enlightening to some.  I did not mean for it to be a brand war on features though as clearly I have bias as an Airtronics pilot.  I just wanted to try and answer the question.  I'm happy to discuss further off list, or answer other specific questions on or off list, but for now I'm going to bow out of this thread.

Thanks,

Mark

-----Original Message-----
From: nsrca-discussion-bounces at lists.nsrca.org [mailto:nsrca-discussion-bounces at lists.nsrca.org] On Behalf Of Anthony Frackowiak
Sent: Tuesday, August 25, 2009 12:55 AM
To: General pattern discussion
Subject: Re: [NSRCA-discussion] SD-10

I'll try to explain this as I understand it. With a 2048 system you  
get 2048 TX step outputs through the entire stick travel only if the  
electronic throws in the TX are set to Max. Say that's 150% Travel at  
100% Dual Rate. Those numbers vary amongst the brands and models but  
this is a start. Any reduction from there will reduce the 2048 by some  
percentage. Let's say Travel is set to 100% and Dual Rate remains  
100%. You have just reduced the number of steps by 33% so now you're  
down to around 1350. If you reduce the Dual Rate to 80% from there  
you're now flying on 1080. If this started out as a 1024 you would end  
up at 540. I guarantee you'll feel the difference between 1080 and 540.

I always try to maximize the electronic throws and set the mechanicals  
to get the surface travel I desire. But some things are inevitable. To  
get the model to properly break in a spin takes a lot of elevator  
deflection, so I end up flying with DR set to 50% or so for normal  
flying. The higher 2048 system makes a difference with this sort of an  
example. And I do believe the better servos will respond to this fine  
of a command.

Hope this isn't too confusing.

TonyF



On Aug 24, 2009, at 8:14 PM, Atwood, Mark wrote:

> The resolution is for the full throw of the servo. If you bump your  
> endpoints up to 150% and have 135deg of travel, you have 512 / 1024  
> or 2048 individual points in between.  At a normal 100% that's  
> across 90degs of throw.  If you dial your endpoints down, it's  
> spread across a smaller arc.
>
> We're talking about extremely fine movement even at 512 resolution.
>
> What's nice is that all the components of the Accuracy "system" are  
> improving together.  The pots in the new radios (all brands) are  
> significantly more accurate, the digital servos of all brands are  
> significantly more accurate, and our linkage systems continue to  
> improve as well.  My argument was simply that after 1024 (and  
> probably lower actually), the resolution is probably not the "weak  
> link" in the accuracy chain.  We have too much slop elsewhere.  Even  
> a tight servo has some lash, as do the connections to the servo and  
> even movement of the control surface on the hingeline.
>
> Admittedly though, On a LARGE control surface like a 40% IMAC  
> Rudder, you can still see the surface "Step" if you move it slowly  
> and watch carefully even with 1024.  So to say that 2048 is not  
> smoother would be foolish.  But on our pattern planes and the throws  
> we use, the other sources of error take resolution off the critical  
> path to more accuracy.
>
> So to slightly change the topic, how do we get spec's on the  
> resolution/accuracy of the servos?  Does anyone publish that? (I'm  
> pretty sure Atx doesn't).  That would be as important than just  
> about any other spec I would think.
>
> -----Original Message-----
> From: nsrca-discussion-bounces at lists.nsrca.org [mailto:nsrca-discussion-bounces at lists.nsrca.org 
> ] On Behalf Of Lisa n Larry
> Sent: Monday, August 24, 2009 10:31 PM
> To: 'General pattern discussion'
> Subject: Re: [NSRCA-discussion] SD-10
>
> I'm curious...
>
> How many degree of servo throw for 1024/2048 are we talking?
>
> -----Original Message-----
> From: nsrca-discussion-bounces at lists.nsrca.org
> [mailto:nsrca-discussion-bounces at lists.nsrca.org] On Behalf Of  
> Ronald Van
> Putte
> Sent: Monday, August 24, 2009 8:34 PM
> To: General pattern discussion
> Subject: Re: [NSRCA-discussion] SD-10
>
> Geez!  I was happy with 512.
>
> Ron VP
>
> On Aug 24, 2009, at 3:13 PM, mike mueller wrote:
>
>> So a 2048 is smoother than a 1024 by double. I wonder at what
>> point you no longet "feel the difference".?
>> For me it could be as soon as tomorrow. M2
>>
>> --- On Mon, 8/24/09, Jay Marshall <lightfoot at sc.rr.com> wrote:
>>
>>> From: Jay Marshall <lightfoot at sc.rr.com>
>>> Subject: Re: [NSRCA-discussion] SD-10
>>> To: "'General pattern discussion'" <nsrca- 
>>> discussion at lists.nsrca.org>
>>> Date: Monday, August 24, 2009, 3:05 PM
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> Frank, you
>>> are getting your terms mixed
>>> up. The 1024 & 2048 refer to the number of
>>> "steps" from one end
>>> of the stick movement to the other. The speed, known as
>>> latency, is the time it
>>> takes from the instant you change the stick to the time the
>>> servo moves. Actually,
>>> it is the time until the servo receives the command to
>>> move. The two functions
>>> are not dependant upon each other.
>>>
>>>
>>>
>>> As far as
>>> keeping up, that depends on
>>> servo speed. It is possible to have a switch change which
>>> instantly commands a
>>> function from low to high, for example. The time to get
>>> there, however, will
>>> depend on how fast the servo can move.
>>>
>>>
>>>
>>>
>>>
>>> Jay Marshall
>>>
>>>
>>>
>>> -----Original
>>> Message-----
>>>
>>> From:
>>> nsrca-discussion-bounces at lists.nsrca.org
>>> [mailto:nsrca-discussion-bounces at lists.nsrca.org]
>>> On Behalf Of
>>> frank
>>>
>>> Sent: Monday,
>>> August 24,
>>> 2009
>>> 3:30
>>> PM
>>>
>>> To:
>>> 'General pattern
>>> discussion'
>>>
>>> Subject:
>>> [NSRCA-discussion] SD-10
>>>
>>>
>>>
>>> Hi Anthony,
>>> Jim, Mark , and Others
>>> who Responded,
>>>
>>>
>>>
>>> Thanks for
>>> your valuable input.
>>> I've read that this radio's resolution is or is
>>> among the very
>>> fastest; so much so that even digital servos
>>> can't keep up with it.
>>> I understand  that the system is 1024 , but
>>> can't help but
>>> wonder why it isn't 2048. I'm sure
>>> I'll get past my
>>> hangup, but would appreciate everyone's
>>> insight's on that
>>> one.
>>>
>>>
>>>
>>> Thanks,
>>>
>>> Frank
>>> Imbriaco
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> -----Inline Attachment Follows-----
>>>
>>> _______________________________________________
>>> NSRCA-discussion mailing list
>>> NSRCA-discussion at lists.nsrca.org
>>> http://lists.nsrca.org/mailman/listinfo/nsrca-discussion
>>
>>
>>
>> _______________________________________________
>> NSRCA-discussion mailing list
>> NSRCA-discussion at lists.nsrca.org
>> http://lists.nsrca.org/mailman/listinfo/nsrca-discussion
>
> _______________________________________________
> NSRCA-discussion mailing list
> NSRCA-discussion at lists.nsrca.org
> http://lists.nsrca.org/mailman/listinfo/nsrca-discussion
>
>
>
> _______________________________________________
> NSRCA-discussion mailing list
> NSRCA-discussion at lists.nsrca.org
> http://lists.nsrca.org/mailman/listinfo/nsrca-discussion
>
> No virus found in this incoming message.
> Checked by AVG - www.avg.com
> Version: 8.5.392 / Virus Database: 270.13.60/2311 - Release Date:  
> 08/20/09 06:05:00
> _______________________________________________
> NSRCA-discussion mailing list
> NSRCA-discussion at lists.nsrca.org
> http://lists.nsrca.org/mailman/listinfo/nsrca-discussion

_______________________________________________
NSRCA-discussion mailing list
NSRCA-discussion at lists.nsrca.org
http://lists.nsrca.org/mailman/listinfo/nsrca-discussion

No virus found in this incoming message.
Checked by AVG - www.avg.com 
Version: 8.5.392 / Virus Database: 270.13.60/2311 - Release Date: 08/20/09 06:05:00
_______________________________________________
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/20090825/9f5e4116/attachment.html>


More information about the NSRCA-discussion mailing list