rrTV-PHOTO   New HD TV
HOME   rrTV-PHOTO   GALLERIES   MY GALLERY   HELP-FAQ
myHOME PM pmRR MEMBERS 348 ONLINE 20 EVENTS SEARCH REGISTER  START HERE
 
54 pages [ <<    <     38      39     ( 40 )     41      42     NEXT    >> ]85452 viewsPOST REPLY
Model Rectifier Corp . Heli Wholesaler . 3D Heli Depot

.
.
Radio - Servo - Gyro - Gov - Batt > TX/RX eCCPM Latency Test Results
 
 
Yug
rrProfessor
Location: UK. Herts

Oops, looks like I opened a can of worms with the servo resolution thingy.
With regards to servo resolution, there are so many ways of approaching the PID algorithm given the motor topology and specification targets with respect to quiescent, max current draw, dead zone parameters and suchlike. Each manufacturer has their own approach but as a general rule, I tend to go with the averaging approach with regards to the ability of a servo to achieve an apparent target position. The nature of heli control tends to be a dynamic issue where inputs are never static but are always changing. This implies a greater importance on the servos ability to faithfully track inputs where interim transitions are smooth and target positions are met with a degree of an averging function for a given static input.

Vegetable rights and Peace
02-25-2008 02:04 AM
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

The average analog servo doesn't get that fancy.

The input pulse is compared to the pulse generated from the feedback pot position and a pulse is generated and fed to the motor. I say a pulse because switching on the motor current and keeping it on until the next pulse will not work because it would allow considerable over-shoot.

The size of the difference between the input pulse and the feedback pulse generates a variable pulse length to the motor. Minor difference, small pulse length (less than 20ms). Major difference, longer pulse length (still less than 20ms). Timings get empirically tweaked until it "seems right".
02-25-2008 02:12 AM
HOMEPAGE  
 
 
Yug
rrProfessor
Location: UK. Herts

got any schematics of servos ?

Vegetable rights and Peace
02-25-2008 02:22 AM
 
 
GimbalFan
Elite Veteran
Location: Copter County, Nv

Quote 
No one can tell me that I have not experienced smoother servo motion with a 2048 radio.
John, in what way have you experienced this difference? Is it while testing on the bench, or can you truly sense a difference in the sky? If while airborn, in what sort of maneuvers is it most apparent?

op-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-t
02-25-2008 09:16 AM
 
 
derek533
Senior Heliman
Location: Edmond, Oklahoma

Does anyone know how the 9C stacks up when using a Fasst module? I would guess it would be very similar to a regular non-PCM FM setup. Am I correct?
02-25-2008 03:12 PM
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

> Does anyone know how the 9C stacks up when using a Fasst module?

My guess is around the DX6/AR6000 numbers. Unfortunately, the 9C remains on the slow end even in PPM mode. Then add latency from the FASST system. Just a guess, though. It will have better CCPM then with PCM1024, though.

> John, in what way have you experienced this difference? Is it while
> testing on the bench ...

Admittedly mostly on the bench while testing radios and setting up helis. Since my personal ownership of 2048 radios has been limited (14 MZ and 12Z) time wise, the vast majority of flying I've done on 2048 has been other people's helis on which I have no control over setup and all the other subtleties. Nor have direct comparisons been possible.

> or can you truly sense a difference in the sky?

When I flew the Vibe I had with a CycLock and the radio setup jacked up to use the full 1024 resolution on each control (ail, ele, and collective) there was definitely a feeling of finer control compared to flying the same heli with a normal CCPM setup over a 1024 radio link (DX7 with AR9000). But that was more than just 1024 versus 2048.

Until now with the X9303, I am not aware of any other way to directly and honestly compare 1024 versus 2048. Since I don't own an X9303, I can't do that myself. I just received a PM from someone who states he definitely feels a difference between the two (AR7000 and AR9000) on the same heli while changing nothing else.

BTW, I am not saying 2048 is necessary for everyone nor am I saying all helis will benefit from 2048 due to perhaps mediocre servos, slop, or just poor setup. What I am saying, is that saying 2048 makes no difference and has no benefit over 1024 is incorrect.

- John

MSH Protos
02-25-2008 03:32 PM
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

BTW, JR states that their DS series servos have 5,900 steps of resolution over their 120 deg operating range.

- John

MSH Protos
02-25-2008 05:01 PM
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Makes me wonder what kind of A/D converter they have in that servo, 13 bit ?
02-25-2008 05:16 PM
HOMEPAGE  
 
 
GimbalFan
Elite Veteran
Location: Copter County, Nv

Quote 
saying 2048 makes no difference and has no benefit over 1024 is incorrect.
From a purely technical aspect one must naturally agree with this.

What I'm having a hard time picturing in my mind though, is that in order to feel any significant difference between 1024 and 2048 increments, it would require that there be zero slack in all of the servo gears, servo bearings, link ends, linkage pivot bearings, swash bearing and head bearings.

It strikes me that even the slightest slack at any one of these points would negate the perceivability of a difference -- and this doesn't even take into consideration the intended flexability of the dampeners.

Supposing we could achieve zero slack everywhere in the mechanical 'chain of command' (is this even possible without inducing a slight but unacceptable level of binding?), how could a human -- with the skeletal/musculature/nervous system specs we operate under -- discern the difference between 1/570" and 1/1140" of stick movement?

op-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-t
02-25-2008 05:25 PM
 
 
MrMel
Elite Veteran
Location: Lidingo, Sweden

Slop in the mechanics only become noticable when you change the load, as long as you have load in one direction its pretty non-existing.

I know when I started with Align T-rex 450 version 1, it was "slop defined", twichy around center, otherwise fine.
02-25-2008 06:01 PM
 
 
GimbalFan
Elite Veteran
Location: Copter County, Nv

Quote 
Slop in the mechanics only become noticable when you change the load
This makes sense, and it does address part of the puzzle.
Quote 
twichy around center, otherwise fine.
Ain't that the truth.

op-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-t
02-25-2008 06:06 PM
 
 
Four Stroker
Veteran
Location: Atlanta

JR states that their DS series servos have 5,900 steps

That sounds like a challege for you John ! That's a much bolder statement than FASST is as good as G3. And the test results would settle this motor pole question once and for all time.
02-25-2008 06:10 PM
 
 
MrMel
Elite Veteran
Location: Lidingo, Sweden

Btw, You talk about 3pole motors, you know that first number in Futaba's servo number is the type of motor.

3 = 3 pole
5 = 5 pole
7 = 7 pole
9 = coreless

So, it can well be that old 3xxx type servos dont have good resolution, but thats why coreless was invented.

Now Futaba says Brushless have even higher resolution though.
02-25-2008 06:40 PM
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Yug, Kraft, almost 40 years old but still the same idea,

02-25-2008 08:54 PM
HOMEPAGE  
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

> how could a human -- with the skeletal/musculature/nervous system
> specs we operate under -- discern the difference between 1/570" and
> 1/1140" of stick movement?

Did you account for the fact that you do not use nearly the full radio link resolution for a CCPM setup? You may only use a 1/4 to a 1/3 of the spec'd radio resolution for any given input function (aileron, elevator, and collective).

For example, on the Vibe I had, full left to full right aileron only used 366 steps of the radios 1024 possible steps. So that 366 steps is from full left stick to full right stick. That's why using the CycLock where I had 1024 steps from full left stick to full right stick made such a difference.

- John

MSH Protos
02-25-2008 11:37 PM
 
 
GimbalFan
Elite Veteran
Location: Copter County, Nv

Quote 
Did you account for the fact that you do not use nearly the full radio link resolution for a CCPM setup?
No I didn't. I understand some (but not all) of the circumstances which reduce actual resolution from potential in a given setup. The degree and percentage of difference you describe there is more like what I imagine to be perceptible.

Do you think that mechanical slack in a linkage system can be minimized to the point that -- in a setup in which a radio/servo system's full resolution potential is achieved -- that (were it possible to switch between 1024 and 2048 mid-flight) a competent flier would really be able to feel a difference?

op-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-t
02-26-2008 12:12 AM
 
 
Yug
rrProfessor
Location: UK. Herts

AirWolf - no PID ? A servo will never achieve optimium performance without a PID algorithm either implemented digtally or in analog.

Gimbal - with regards to the 'sensitivity' of ones fingers, I liken this to the kind of sensitivity required by proffessional musicians, particaularly violinists, cellists etc; where an exceedingly precise and intimate tactile relationship is developed between finger and string, where the sensitivity exceeds what is overtly measurable. I've often likened flying a heli to playing an instrument; at the upper echlons, a similar level of intimacy, sensitivity, objectivity and a 'connection' that enables a fluid and expressive performance.

Vegetable rights and Peace
02-26-2008 01:01 AM
 
 
GimbalFan
Elite Veteran
Location: Copter County, Nv

Quote 
I liken this to the kind of sensitivity required by proffessional musicians, particaularly violinists, cellists etc... I've often likened flying a heli to playing an instrument
Couldn't agree more. I'll go a step further and speculate that it's the delightfully enticing mix of left-brain and right-brain activities which make RC in general and helis in particular so goose-bumpiliciously addictive.

op-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-thwõp-t
02-26-2008 01:04 AM
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Quote 
no PID
Like I said, they're not that fancy.
02-26-2008 01:08 AM
HOMEPAGE  
 
 
Yug
rrProfessor
Location: UK. Herts

Left and right brain oops, mine have amalgamated into a confused mush.

AirWolf, I would very supprised is modern servos aren't more advanced than your schematic when considering the likes of the Hitech programmable servos. For me, this implies that they incorporate a degree of 'digital intelligence' in so much as a hybrid algorithm is incorporated which allows optimisation of performance. With this, it follows that PIDs with null windows are employed although I'm still sceptical over the ultimate resolution of a servo given the basics of gear ratios and the maximum resolution of motor step positons in balance with the maximum acceptible quiescent current draws.

Vegetable rights and Peace
02-26-2008 01:29 AM
 
 
54 pages [ <<    <     38      39     ( 40 )     41      42     NEXT    >> ]85452 viewsPOST REPLY
JR-Spektrum . Gyro Hobbies . E-flite

.
.
Radio - Servo - Gyro - Gov - Batt > TX/RX eCCPM Latency Test Results
 PRINT TOPIC Advertisers 

Subscribe to This Topic

Saturday, November 22 - 12:21 pm - Copyright © 2000 - 2008 runryder.com | email | link to rr | runryder needs cookie