rrTV-PHOTO   New HD TV
HOME   rrTV-PHOTO   GALLERIES   MY GALLERY   HELP-FAQ
myHOME PM pmRR MEMBERS 757 ONLINE 28 EVENTS SEARCH REGISTER  START HERE
 
51 pages [ <<    <     6      7     ( 8 )     9      10     NEXT    >> ]79344 viewsPOST REPLY
3D Heli Depot . JR-Spektrum . Gyro Hobbies

.
.
Radio - Servo - Gyro - Gov - Batt > TX/RX eCCPM Latency Test Results
 
 
Harris
Senior Heliman
Location: A Greek in Cyprus

Quote 
> How about non eCCPM mode? Is there a slew rate limitation there too?

I just tested it and no, there is no slew rate limiter on collective in normal swash mode. Added latency data for the collective channel in normal swash mode.

Interesting! Care to speculate why? I know for fact that eCCPM is very math-intensive (lots of high precision real number arithmetic) so I assume the CPU simply can't keep up but I have no hard data to support this.
04-16-2007 Over year old.
HOMEPAGE  
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Quote 
I know for fact that eCCPM is very math-intensive (lots of high precision real number arithmetic) so I assume the CPU simply can't keep up but I have no hard data to support this.
Actually, no
Looking at 120º CCPM, to get the elevator values, send the data directly to the aft servo and divide by 2 (cos 120º) for the data to the two front outboard servos. To get the aileron values, just multiply by 0.866 and send the data out to those two servos. For collective data, just add or subtract the same value from the 3 servos.

To multiply by 0.866 (sin 120º), you could just subtract 1/8 of the value from itself and be within 1%. 1/8 is a simple shift of 3 bits to the right.

Shift 4 more bits and also subtract that and you're within 0.1%
04-16-2007 Over year old.
HOMEPAGE  
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

0.766 (cos 140º) instead of 0.5 for the elevator and 0.643 (sin 140º) instead of 0.866 for the ailerons.

In the end it's no fancy math, just bit shuffeling.
04-16-2007 Over year old.
HOMEPAGE  
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

If the front and rear servos were equal distance to the front and to the rear from the main shaft, they would be 180º

I think the purpose of 140º is to account for the time it takes for a servo to get to it's target position. With the front-off-to-the-side servos moving 1/2 as much distance as the rear servo, they get there faster with 120º CCPM and cause a bit of collective jump.

With 140º, the jump is less.
But this only matters with real fast stick movements.
04-16-2007 Over year old.
HOMEPAGE  
 
 
village_idiot
Senior Heliman
Location: Land of confusion

Can't wait to see if there are any changes as these new 2.4Ghz modules start shipping. One comment on RCGroups was that his 8U with the XPS system was now faster.
04-16-2007 Over year old.
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

The radius from the main shaft to the servo atachment points are the same for 120 and 140. The for/aft distances are not.

Given the same radius and the same for/aft distances, the result would be 180º

Draw it out to scale and see what you get.
04-16-2007 Over year old.
HOMEPAGE  
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

> Care to speculate why?

My only speculation as to why is an attempt to keep the swash movement purer given the up to three frame spread in when data gets to the servos. The slower the servos move, the less swashplate tilt a given time delta in servo command will cause.

BTW, the elevator and aileron inputs are also slew rate limited in eCCPM mode.

> The radius from the main shaft to the servo atachment points are
> the same for 120 and 140.

The radii do NOT have to be the same and thus the fore/aft distance relationship can be played with. I entire intent of a good "140 deg eCCPM" implementation is to have equal fore/aft distance of the balls. The "140 deg" in "140 deg eCCPM" is just a way to designate the difference from 120 deg eCCPM; the actual angle used is not 140 degs.

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Before this 120 v 140 debate goes any further, I think a clear mechanical definition of 140 is in order.
04-16-2007 Over year old.
HOMEPAGE  
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

> I think a clear mechanical definition of 140 is in order.

The mechanical definition is quite simple... Make the fore/aft distance of all three control points the same. That's it.

That makes the math extremely simple since all servo movements are then equal just in the same or opposite direction as required for a given desired swash motion.

The left/right seperation of the two balls is irrelevant and does not change the math required. The seperation would, however, influence the perceived match of elevator response speed versus aileron response speed. Notice I said "response speed", not maximum cyclic rate.

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
DS 8717
rrProfessor
Location: Here wishing i was somewhere else

The only problem i can see with 140 is if the ele servo does not have enough torque,it could be slower than the ail and pitch servo because they have twice the availble torque.DOUG
04-16-2007 Over year old.
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Given the for/aft distance is the same and the angle is 140º,
that leaves only 80º between the "aileron" servos,
and requires that the radius to the "aileron" servos be 30.54% greater than the radius to the center aft servo.

Now if you use 135º,
You can have the same for/aft distance AND the same side to side distance with the radius to the two "aileron" servos being 41.4% greater.
- - - Which sounds like what 140º wanted to be in the first place.
04-16-2007 Over year old.
HOMEPAGE  
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

Doug,
Yep, that's the one problem of "140" deg eCCPM. Because the moment arms are the same for all three servos for collective inputs, you now have more force available on the fore or aft "two-ball" "side" of the swash.

OK, enough eCCPM geometry talk. This thread is about latency which does have an effect on how a eCCPM swash ends up moving, but no geometry talk is necessary.

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

All right, one last eCCPM geometry item. Here is a picture of the Vibe in a "140º CCPM" configuration. Notice that the radii are definitely not the same. http://helicopterosrc.com/vibe/vibe5.jpg

Quote from Vibe manual: "CCPM 140 places the ball links on the swashplate 140º back from the forward ball, then the rear balls are extended, placing them the same fore-to-aft distance from the center of the main shaft as the front ball."

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

Quote 
Yep, that's the one problem of "140" deg eCCPM. Because the moment arms are the same for all three servos for collective inputs, you know have twice as much force available on one "side" of the swash.
? ? ? ?
04-16-2007 Over year old.
HOMEPAGE  
 
 
Chuckie
Key Veteran
Location: Crofton Maryland, MHA member

Quote 
Yep, that's the one problem of "140" deg eCCPM. Because the moment arms are the same for all three servos for collective inputs, you know have twice as much force available on one "side" of the swash.

Quote 
OK, enough eCCPM geometry talk. This thread is about latency which does have an effect on how a eCCPM swash ends up moving, but no geometry talk is necessary.

So there is latency in the elevator movement in 140 CCMP if there is high loading on the swashplate

Nice

120 and 140 CCPM would have the same problem.

Chuckie

Please stand by for faster service!
04-16-2007 Over year old.
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

>
Quote 
Yep, that's the one problem of "140" deg eCCPM. Because the moment arms are the same for all three servos for collective inputs, you now have more force available on the fore or aft "two-ball" "side" of the swash.

I'm thinking this one over. The explanation and math made sense at the time I was exposed to it. I don't think "twice" is quite right.

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

If the force is applied further out, there will be less resistance on the servo causing the servo closer in (smaller radius) to carry more of the load.

Presumably, with less load, servo travel time is faster. The servos further out will move faster than the servo closer in (radius).


The 140 arangement is better but still not all there.
04-16-2007 Over year old.
HOMEPAGE  
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

Now you're thinking AirWolf. 120 deg is balanced for collective but causes elevator input issues. "140 deg" fixes the elevator input issue but causes imbalance in collective inputs. After doing the math, twice is indeed correct.

Here's an explanation...
Assume flybarless head, "center" ball is on the aft side of the swash, equal length servo arms, balls for blade grip links are aligned with the longitudinal axis, looking at the system from the side perpendicular to the longitudinal axis, and equal fore/aft distance of the swash control points.

The swashplate ball is a pivot point. For a collective input, the force required on the two links going to the blade grips would be the same and they are the same distance from the pivot point. In order for the swash to not tilt, the total torque on each "side" (fore and aft) of the pivot point must be the same. Since you have two front servos the same distance out as the third servo on the rear, the two front servos must each supply half the force that the rear servo must.

Thus, the front servos are more lightly loaded and would, in theory, move faster than the twice as heavily loaded rear servo.



I am not saying that this is a problem which actually shows itself in real flight. Given sufficiently strong servos, it would be minimized perhaps to the point of a complete non-issue.

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
AirWolfRC
rrProfessor
Location: 42½ N, 83½ W

So let's go back to 120º and have a mix in the heli from the elevator servo to the "aileron" servos for speed only during elevator commands and all is well
04-16-2007 Over year old.
HOMEPAGE  
 
 
JKos
Elite Veteran
Location: City of California in the state of Maryland

> So let's go back to 120º and have a mix in the heli from the
> elevator servo to the "aileron" servos for speed only and all is
> well

As has been done in a few radios including the venerable Stylus. Why this has not been implemented in more radios is sad.

- John

Protos -- Logo 10
04-16-2007 Over year old.
 
 
51 pages [ <<    <     6      7     ( 8 )     9      10     NEXT    >> ]79344 viewsPOST REPLY
E-flite . Next D . Fast Lad Performance

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

Subscribe to This Topic

Wednesday, October 8 - 1:47 am - Copyright © 2000 - 2008 runryder.com | email | link to rr | runryder needs cookie