Zor

Hi Ho Gals and guys,

From discussing "resoluion" it is quite nomal to elaborate into other aspects of the communication link to our model aircrafts.

As mentioned in previous posting, I opened my JR XP7202 and took a few pics.

I am attaching some here to give an idea to those interested, what a typical transmitter looks like inside.
I believe these transmitters are all quite similar.

At top center is the RF (Radio Frequeny) generating board and the litlle white wire at the bottom left of that board is going to the base of the antenna.

The spring leaf (strip) riding on he indentation segment is visible at the right stick structure and that ridged segment also exist at the left without the spring. No clicking on the elevator whish is at the left in the pic since we are looking from the back.

The PCB (Printed Circuit Board) at the bottom is where all the modulating signal is generated .
Looking at the RF board the connector at middle left has 3 wires. The middle yellow wire is the connetion that brings the modulating signal to the RF and is shown in one picture on the face of the oscilloscope. It is not easy to try to synchronize this wave form which is pulses of different timing .

I have a Spectrum DX7 which is nearly identically built. I did not open it but it is no doubt very similar but for the RF board generating the 2.4 Ghz instead of the 72 Mhz.

There is no Xtal in the 7202 since it is "synthesized" . We can select any radio channel by programmng except the three lower frequency RF channel. They would be the image frequency on the three upper channels.

Any questions are welcome without any guarantee of a proper answer from me. If I do not know I will simply say "I do not know".

I believe we have pretty well reached our own conclusions on the topic of "resolution".

There is so many aspects of this fine hobby to keep our minds and our leisure times occupied in a happy way.

Regads to all from Zor.

Flying Geezer

My Feedback: (14)

I don't think radio resolution has anything to do with servo resolution. I think is the rate of speed that information, (modulation - stick inputs) are transmitted in a given period of time. The 2048 would then be twice as fast, ergo 11ms instead of 22ms.

Think of baud rate on computer modems. My first modem had a baud rate of 300. The next one was 1200, we thought it couldn't get much better. Technology advanced rapidly to 48000 + on dial-up, and now with broadband we hit up to 20 mHz.

I have a 2048 system and I'm no 3D pilot, but I felt the difference immediately.

Just my opinion.

Zor

Hello Flying Geezer,

My present understanding is pretty well in agreement with the way I understand yours.

The 1024 and 2048 resolution has to do with the rate of sending commands to the receiver and passed along to the servos.

The way I understand at this moment the chain of events, it does not affect the servo speed of action unless a whole new bunch of servos exist to take advantage of the higher transmitted resolution.

I seetwo factors (two characteristics or properties) here involved ___
1 The commonly known speed of servo rotation. Specs are pulished and can be consulted.
2 The "latency" which I understand as being the elapsed time between the transmission of a command (like a position change) and the time the servo is being told to follow suit. The interpretation time within the servo would be part of the latency the way I understand things.

I have not seen anything related to the above two factors else than the typical advertising claims without factual information. If it exist I did not come across it. I wonder what is affected by the 11 milisecond vs the 22 else than a shorter frame.

I even wonder if a shorter frame that might have less infomation might need a longer interpretation on board the aircraft. I am trying to find as much as I can "how all this work" but it is not easy for me to find info and data.

Thanks for coming in. I thought initially that this thread would die within just a few postings.
It is nice to see some people interested in this aspect of the hobby.

Best regards de Zor.

Edited to change spelling. The word positon became position.

BarracudaHockey

My Feedback: (11)

Nope, there's 2048 discreet steps in the stick encoders.

To wit, there are 2048 resolution recievers that operate at 22ms or 11ms latency. Infact, the only radio currently capable of 11ms latency is the DX7SE (not sure about the DX8 as I haven't played with one yet)

JNorton

My Feedback: (2)

BarracudaHockey,

Airtronics SD-10G:
Min: 7.36ms
Max: 16.3ms
Ave: 11.83ms

I don't think that's too shabby. I like the fact that it is FHSS. I'm not the pilot you are and never will be but for a sport pilot I'm seriously considering this transmitter.
John

Zor

Hi BarracudaHockey and all readers,

In order to have 2048 resolution the DC voltage from the potentiometers that are rotated by the sticks have to have at least that many DC votage variatons for the ADC converter(s) to produce that resolution.

It is difficult to visualize a potentiometer contact that would touch the resistive deposit in an arc contact that is as small as 1 / 8192 of a circle. These potentiometersappear to be only about 1/2" diameter (at the most 3/4").

It is easy to convert any fixed DC potential to any number of pulses and I am now wondering if we may have 2048 resolution in increments of much smaller available output. That is difficult to put in words.

The DC voltage from the potentometers is about1 volt to 8 volts. and the neural at 4.5 volts. As the stick is moved and the voltage varies between th 1 and the 8 volts it is possible to have a DC at any one instant controlling a 2048 capable ADC converter but I cannot see the DC out of the potentiometers being capableof giving out ____
as an example right now I have 2.5000 volts and the next moment I can have a voltage of 2.503418 or 2.496582 volts or even a smaller vaiation.

At this stage I believe that these transmitters have the capability of 2048 resolution out of their ADC converters and the number of digital bitstotransmit digitally that resolution but it seems to me that the ADC need that degreee of DC resolution to actually produce that many increments.

If I understand correctly, the 11 digital bits are controlling the width of the pulses but I cannot get out of my head that the final resolution meaning the capabiliy of having a servo move by a rotation of 0.043945 degrees is actually achieved.

I can only see the final practical resolution as being determined by the resolution of the DC voltage out of the stick potentometers where we have to have the possibility of picking a voltage variation as small as 0.003418 volt with a mechanical cursor moving over a deposited resistive surface.

Then we have a similar situation in the servo(s). In the next couple of days I will put an analog guage on a servo arm and see how small a movement I can get reliably and will calculate the angle of rotation that occurs. If anyone is interested you can come back to this thread and look at my next posting.

Any different reasoning that makes better sense would be much appreciated.

I am confused about the 22 ms and the 11 ms being referred to as latency.
I stand to learn every day.I thought (and still do) that latency was referring to a delay in response and that the 22 ms and 11 ms were frames.

Hopfully some day soon I may understant all this properly.
Meanwhile I can still fly my model aiplanes.

Regards,

Zor

Zor

Hello Gals and Guys,

This posting is aimed at interested people and is looking at resolution on a practical viewpoint.

Involved here is a Spektrum DX7, a R7000 Spectrum receiver and a Spektrum DS821 servo.

I am simply trying to give myself an idea of the kind of resolution I have.

I made a setup as seen in the attached pictures.
I assured myself that the guage is mounted solidly enoughso that the servo motion does not affect the gauge mounting. It is a crude but reliable set up.

Below some of the results I observed ___

Centering ___the servo comes back to center within +/- 0.001"
Smallest increment I could get using the transmitter stick +/- 0.002 sometime 0.003 but always at or in between these values.
Trim ___the digital trim on a single beep gave me 0.003" to 0.004" occasionally 0.005" .

Now I have to interpret these figures in terms of resolution.
When resolution is quoted as being 512 or 1024 or 2048 I have to assume (have to understand) that it is within the servo normal maximum range of travel

I will use +/- 60* (the asteric will mean degrees) as the normal range of rotation of the servo.

In my tests the guage is accuately enough 1/2" from servo axis of rotation.
The cicumference is 0.500 X (multiplied by) pi (3.14159265) or circ = 1.57079633 inch.

Yes I am using very many decimals but I am having fun and hurting no one.
Now ___60* would be 1.57079633 * 60 / 360 (or 1/6 of circ.) and that is 0.26178 . . . . .
and that would be half of either 512 or 1024 or 2048 of the resolution.

Since I am measuring the motion at 90* with the servo horn and only playing a minimum rotation I can say that the first increment of motion would be ___
for 512 ___0.26178 / 512 = 0.00051133
for 1024 __ 0.26178 / 1024 = 0.00025566
for 2048 __0.26178 / 2048 = 0.00012783

When I compare the calculated values in the above 3 lines with the practical results I could measure, I am very far from being able to use such resolutions in practice.

Doing further figurations ___
I understand (right or wrong) that the DX7 has a resolution of 1024.
I figure that to be for the full 120* of servo rotation ( plus and minus 60*).
So going one way the first increment would be 0.00051 but I cannot get less than 0.002" to 0.003".

I have to conclude that with my rig (as described at the beginning of this post) my practical resolution is 0.002 / 0.00051 = 3.92 or 4 times worst (not as good).
and 0.003 / 0.00051 = 5.88 or 6 times worst .

I am not criticising the equipment claims.
I am trying to see what it does for me on a practical viewoint and in the reality of my setup.

The measures I read were by very delicately moving the transmitter stick to get only the very fist signs of servo motions which of course are never used in practical flying.

If you have read up to here, you are as "nutz" as I am ___Hee! Hee!
I do think however that it gives me a pretty clear outline of what I am playing with.

Any decent comments are welcome and derogatory postings by specific individuals are not welcome.
I regret having to write that line above but it has come to that unfortuantely.

Best to everyone from Zor.

JNorton

My Feedback: (2)

I like the fact that a $30.00 3 pole digital sport servo centers to within +/-.001" using an fairly inexpensive transmitter - receiver. That is pretty impressive.
John

Flying Geezer

My Feedback: (14)

The signal is transmitted and received as a specific number, (digital). After the receiver gets the number, it must convert it to a voltage pulse thru a D/A converter and correctly time the length of the pulse. Consequently, you are bound to lose some of the resolution in the conversion, not to mention the accuracy of the servo.

I would conclude that you will not get exactly 1024 or 2048 precisely in the servo movement, however it is a substantial improvement in the way my stick response feels. Further, with the 2048 resolution systems I have used, the response from the stick, (not necessarily the accuracy) is substantially faster than the 1028 systems. I don't move the stick and then see the response. I see the response in real time with the motion of the stick.

Keep in mind that I have flown 1024 on FM and PCM at 72 mHz and the 2048 on 2.4 gHz.

An acquaintace of mine has owned a JR 10X with a 2.4 module and a Futaba 12FG dedicated 2.4 and insists it is much faster.

Zor

Hi Flying Geezer,

Thanks for the contributed information.

I initiated this thread trying to find more about the relationship of the transmitter with the end results of servos response.

I first considered the resolution and titled the thread with that word.
I understood that "latency" is referring to any delay involved and still think so.
No one corrected me from saying so previously.

My measurement experiment obviously had to do only with resolution and not latency which is the main topic of your posting if I read you correctly.

I am glad your participation is very valuable as (again if I am correct about the meaning of latency) the speed of response of the servo to a transmitter command or variation of a command might even be more important than the resolution.

It might be logical to think that the manufacturer going to 2048 may have used this method finding out that it increase the reaction speed.

Thus many people may feel the same way you do.
Now to do some thinking how the application of a higher resolution reduces the latency (delay) time.

Always something fascinating to pursue.

Thanks again and best regards,

Zor

Zor

Flying Geezer

Note that I am always keeping the beginner in mind and that perhaps is not you.

I am not arguing about what you wrote in your postings. I am doing some further thinking.

The model response is to the air stream forces as the commands move the control surfaces and not the model itself.

Even if we assume that the control surfaces move instantly with the transmitter control sticks, we still have some mass of the airplane and some inertia. The control surfaces create rotational forces that depend on the amount of surfaces rotation and the airspeed. There is some observable delays in the model reaction even if very short in duration. The heavier a model is the longer timeisneeded to show an observable change in attitude of trajectory.

I think there is some advantage in a slower response of the model that helps in smooth flying.
Of course we are talking about fractions of a second and I am not using the transmitter sticks as ON / OFF switches. In fact I am using some exponential on all three axis of rotation to make the sticks less sensitive around their neutral positions.

We may consider two main styles of flying ___
1 Controlling by anticipation and making the model do what we anticipate; what we wish the model to do. It helps to have a model well balanced and with some degree of stability/ Neutral stability is fine for aerobatic maneuvers. General sport flying bemefits with some degree of positive stability.
2 Controlling by reaction and constantly trying to correct what the model is doing on its own. That can become a nightmare if the model is poorly balanced or has negative stability. That is often the problem of a newcomer that did not balance his model adequatly or used an improper method of setting the CG location. Ya . . . the CG is XXXX distance from the leading edge of the wing at the root just where it shows on the drawing (instructions). If a proper method of deteminign the CG had been used it would have shown the CG 1/2" or 3/4" out of recommendations.

As we control the model we are moving the control surfaces to create rotational forces around the three axis pitch, yaw and roll. The model is "accelerated" into a rotation and will contimue to rotate if not stopped by using opposite control to stop the rotation.

An example is best observed in a turn (that we do a lot of to keep the model in range and in sight) .
A turn is initiated ___the model rolls and yaw and need some pitch to generate the turning force and maintain altitude. The wing outside the turn travels faster than the inner wing andproduce more lift, thus an overbanking force that needs to be stopped at the banking degree desired.

If this is not "anticipated" (see 1 above) the bank becomes nearly 90 degrees and the model is slipping with considerable loss of altitude. That is so commonly observed at the flyng field.

Who cares? Lots of power on my plane, I can climb back straight up vertical if necessary.(more thrust than weight).and that is fine if that is the style of flying you enjoy.

It seems to me that some of the art (is it an art ?) of flying is lost or has never been acquired by so many fliers.

Nothing wrong with that. We do what we enjoy best and do it within our capabilities.
It just seem to me that mehod 1 above brings more enjoyment and self satisfacton in controlling the model than method 2 that ends up in fighting the model all the time.

Your choice to pursue whatever you prefer.

Enjoy and anticipate to avoid crashing.

Best 2 U all,

Zor

P.S.: Just posting some of my outlook, aim and goals. No need for some ofyou guys to blast me unless you are one that mainly enjoy doing that and go out of subject just to do that without provocation.