RunRyder RC
WATCH
 2 pages [ <<    <     1     ( 2 )    >    >> ] 1700 views POST REPLY
02-04-2008 08:48 PM  10 years agoPost 21
JKos

rrProfessor

Redondo Beach, CA

My Posts: All  Forum  Topic

> What rotational angle are you stating the servo to be able to
> travel then?

Well, a typical 1.0 to 2.0 ms pulse width range would drive a Futaba servo 1/0.4*45 or 112.5 degs. That sounds about right.

> but from what you have above, I can't see how you got to .002

1/400*45 gives the 0.11 degs per 1 us pulse width change. 0.11/360*2*pi gives the 0.002 inches at 1 inch out.

> What I *DO* see is that if you claim that you get .11 degrees per
> step, then for 2048 steps, the servo must rotate 255 degrees

No, the pulse width resolution for a 2048 system is twice that of a 1024 system, so the servo will have the same total travel. Each step would be about 0.056 degs on the 2048 system (0.5/400*45).

> but considering that you can't try just the latency differences
> and just the resolution differences

This can now easily be done. Try a helicopter with an X9303 and an AR9000 rx then an AR7000 rx (or which ever way around). The only difference will be the resolution.

- John

RR rules!

PM  EMAIL  GALLERY  Attn:RR  Quote
02-04-2008 09:23 PM  10 years agoPost 22
AAKEE

rrApprentice

Boden, Sweden

My Posts: All  Forum  Topic

When using eCCPM the three CCPM servos will be used with very small movements for cyclic(both nick and roll) movement, this (I guess...)is done so simultanious large pitch movements and large cyclic wont get the servo to the end position.

Its a very small portion of the servos normal movement that will be used for a maximum right or left or forward or aft stick position.

As I see it, the resolution when using eCCPM would be lower then if you are using normal mix( the cyclic servo uses all its normal movement for maximum pitch).

So, in my eyes it would be a good idea to get better resulotion due to this factor ? ...but I guess the most of it in this case would be done in the servo? ...I read somewhere that the servos themself wherent so high in resulotion.

If I look at my trex 450 with hitec hs56hb on ccpm, it seems that the resulotion ain´t that good. (just moving pitch or cyclic in small steps and watching servo movement). I have the Dx7+ AR7000.

PM  EMAIL  Attn:RR  Quote
02-04-2008 09:32 PM  10 years agoPost 23
ZAC ATTACK

rrKey Veteran

Hamilton Ontario, CANADA

My Posts: All  Forum  Topic

I am going to purchase a 2.4Ghz X9303 Tx, will it fit in a Futaba Aero Team carrying case? Does the antenna collapse? If not ,can someone tell me the hieght from the bottom to the tip of the TX!!! Sorry for getting off topic boys!

MAAC#77677 Medicated daily for your protection

PM  EMAIL  Attn:RR  Quote
02-05-2008 02:45 AM  10 years agoPost 24
tadawson

rrElite Veteran

Lewisville, TX

My Posts: All  Forum  Topic

1/400*45 gives the 0.11 degs per 1 us pulse width change. 0.11/360*2*pi gives the 0.002 inches at 1 inch out.

> What I *DO* see is that if you claim that you get .11 degrees per
> step, then for 2048 steps, the servo must rotate 255 degrees

No, the pulse width resolution for a 2048 system is twice that of a 1024 system, so the servo will have the same total travel. Each step would be about 0.056 degs on the 2048 system (0.5/400*45).

> but considering that you can't try just the latency differences
> and just the resolution differences

This can now easily be done. Try a helicopter with an X9303 and an AR9000 rx then an AR7000 rx (or which ever way around). The only difference will be the resolution.
Your logic is making my head hurt . . . where is the 1/400*45 coming from? I make the assumption, since the radio is NOT necessarily using PPM internally, that the range of pulse widths from 1000 to 2000 (which you state) corresponds to the 2048 steps of resolution, or about .5 per step. All that, though, at least from the angle I have been looking at this is pretty much irrelevant though.

We have (using your numbers) 112.5 degrees of rotation which equates to the 2048 steps, and which also produces (on a 1" servo arm) 1.96 inches of circular motion. Once again, I won't compensate for just the linear component, being generous. Since this 1.9 inches of motion corresponds to 2048 steps, it's 1.96/2048 for the movement per step, which gives .000958 inches per step. Still an insignificant amount.

Just looking at your numbers, for 112.5 degrees of rotation across 2048 steps, that's simple division, and gives .0549 degrees per step - not .11. Methinks you may have been using 1024 by accident there . . that, or you assume that half the resolution is wasted on the 0 to 1000 us range, which is used. I don't - that would be an idiotic design, since the RX does not output that range, just 1000 to 2000 . . In all reality, the digital value is sent across the RF link, and the RX actually creates the pulse train to the servos, so it only makes sense that the range of values transmitted correspond to the range of pulse timings output - not tp those which cannot be output (0 to 1000 or so . . ).

And once again, your comment about the .11 is confusing . . . for it to be correct, with 2048 steps for full range, 2048*.11 ***MUST*** equal full rotation of the servo, and on that, you are way off . . .

I showed all my numbers and math - it's pretty simple geometry and arithmetic . . . no errors there, or if so, show me where I got it wrong . . . I can't check yours, since numbers seem to pop up from nowhere with no explanation . . .

- Tim

PM  EMAIL  GALLERY  Attn:RR  Quote
02-05-2008 03:19 AM  10 years agoPost 25
JKos

rrProfessor

Redondo Beach, CA

My Posts: All  Forum  Topic

Tim,
Your confusion is that you are mixing my 1024 numbers and 2048 numbers. I clearly said the step size was about 0.11 degs for 1024 and 0.056 degs for 2048.

If you think 0.002 inches of movement on 1024 is no big deal, then that is fine as your opinion. It can be the difference between a level swash and a not level swash.

I will agree that 2048 seems to be all we need at the servo with present day servos. Maybe some day the servos will be better and higher res output from the receiver will be useful.

> I showed all my numbers and math

So did I.

> where is the 1/400*45 coming from?

I'll start over. A Futaba servo, per specs given on ServoCity.com, moves 45 degs for a 400 microsecond change in pulse width. A 1024 system has approximately 1 microsecond output pulse width resolution. Therefore, a 1024 system can move a Futaba servo 1/400 of 45 degs. Thus, 1/400*45 or about 0.11 degs.

A one inch radius circle has a circumference of about 6.283 inches. Thus, a 0.11 deg rotation yields an arc length of 0.11/360*6.283 or about 0.00196 inches.

A 2048 system has approximately 0.5 microsecond output pulse width resolution. Therefore, a 2048 system can move a Futaba servo 0.5/400 of 45 degs. Thus, 0.5/400*45 or about 0.056 degs.

A one inch radius circle has a circumference of about 6.283 inches. Thus, a 0.056 deg rotation yields an arc length of 0.056/360*6.283 or about 0.00098 inches.

Again, I ask you, Tim, have you actually witnessed the difference in motion of a good servo on 2048 versus 1024?

- John

RR rules!

PM  EMAIL  GALLERY  Attn:RR  Quote
02-05-2008 06:07 AM  10 years agoPost 26
seattle_helo

rrKey Veteran

Seattle, WA USA

My Posts: All  Forum  Topic

Fascinating info there, John. Thanks for laying out the math.

nick

PM  EMAIL  HOMEPAGE  GALLERY  Attn:RR  Quote
02-05-2008 12:49 PM  10 years agoPost 27
BarracudaHockey

rrMaster

Jacksonville FL

My Posts: All  Forum  Topic

There's some brilliant people in this hobby, John makes me realize I'm not one of them

Andy
AMA 77227
http://www.jaxrc.com

PM  EMAIL  GALLERY  Attn:RR  Quote
02-05-2008 08:08 PM  10 years agoPost 28
tadawson

rrElite Veteran

Lewisville, TX

My Posts: All  Forum  Topic

Looks like we hit the same numbers, once I corrected to 112 degrees of rotation after all . . . interesting the two different approaches to get to the same thing . . .
If you think 0.002 inches of movement on 1024 is no big deal, then that is fine as your opinion. It can be the difference between a level swash and a not level swash.
I think that the 2048 resolution (about 9/10000" as I calculated it) is far finer that the amount of slop in most linkages that I have ever seen, thus rendering it undetectable. As I said at the outset, for those who take the time to do setups to NASA type standards, they may see it . . . the other 98% of the flying public won't . . .

You also make the assumption that current servos are capable of delivering this level of accuracy . . . what with deadband issues, eventual gear slop, etc. I am not sure that I accept that as valid yet either . . . especially after something has been used for a bit.
Again, I ask you, Tim, have you actually witnessed the difference in motion of a good servo on 2048 versus 1024?
This I can't say I have done, and don't plan to any time soon, based on only resolution . . . I don't throw money at problems that I feel I don't have . . .

- Tim

PM  EMAIL  GALLERY  Attn:RR  Quote
02-05-2008 10:36 PM  10 years agoPost 29
TheSteve

rrNovice

Vancouver, BC, Canada

My Posts: All  Forum  Topic

I'm in favor of the higher resolution myself. An example of why is a really well setup RC car with a transmitter that has 1 microsecond resolution. One notch of trim moves the PPM 1us(as verified with a custom 2.4ghz system that has 250ns resolution) On a long straight the one notch of trim can be seen, on the bench you can hear/feel the servo move that one notch. This is with high end Futaba and Airtronics digital servos. I figure if I can see the 1us change in the steering of an RC car I will feel it in a well setup helicopter once my flying skills match my driving skills. So 500ns resolution is gonna make things just that little bit better, bring on 4096 resolution already!

PM  EMAIL  Attn:RR  Quote
WATCH
 2 pages [ <<    <     1     ( 2 )    >    >> ] 1700 views POST REPLY
 Print TOPIC  Make Suggestion 

 6  Topic Subscribe

Tuesday, December 11 - 12:16 am - Copyright © 2000-2018 RunRyder   EMAILEnable Cookies

The RC discussion world needs to consolidate. RR is now one choice for that. Its software is cutting edge. It hosts on-topic advertising. Help RR increase traffic buy making suggestions, posting in RR's new areas (sites) and by spreading the word.

The RunRyder Difference

• Category system to allow Rep/Vendor postings.
• Classifieds with sold (hidden) category.
• Classifieds with separate view new.
• Answer PMs offsite via email reply.
• Member gallery photos with advanced scripting.
• Gallery photo viewer integrated into postings.
• Highly refined search with advanced back end.
• Hosts its own high end fast response servers.
• Hosts thousands of HD event coverage videos.
• Rewrote entire code base with latest technology.
• No off-topic (annoying) click bait advertising.
Login Here
 New Subscriptions 
 Buddies Online