Oct 27, 2013, 05:26 PM
Quote:
 Originally Posted by theKM differential is capping or the new epa setting is just capping?... the new epa code is still proportional up to the intended endpoint, right?... can it stretch the input for servos that can go 180 degrees just by signal input?...
The pulse is not altered, just capped. So, the end point is just a hard stop (cut off) of the servo pulse. The math is simple. Setting the cap to 1800us is done using this logic:

If Pulse >1800us then Pulse = 1800us.

The signal is not altered to compress or expand the pulse to fit within a range.
 Oct 27, 2013, 05:37 PM Thread OP I spent a few hours today experimenting with the new X10+ servo output code and RFU. I found that I can fly hundreds of feet up and away with the button pressed. In fact, I spent a little bit of time with no antenna on and the button pressed. I was able to taxi to the end of our runway (425 feet) from the middle (so about 200 feet) with no antenna and button pressed. The range of the RFU is incredible. I don't recommend you ever do this, and just because system with an RFU passes the normal range test without an antenna, please don't fly without the antenna! The new servo code worked perfectly, so I will be changing everything to that new code so we can have independent output reversing/centering/epas.
 Oct 27, 2013, 06:05 PM Registered User This is TERRIFIC! news.......
 Oct 27, 2013, 06:10 PM Still Screamin'! Wow Jim!! This is amazing stuff. Almost too much to take in with a single reading All great for the ease set up on new jets for sure! What a great thing to eliminate all the wing wiring nightmare Plus the need to buy color coded servo leads so you don't plug the flaps into the ailerons
Oct 27, 2013, 06:57 PM
↓↘→ + (punch)
Quote:
 Originally Posted by JimDrew The pulse is not altered, just capped. So, the end point is just a hard stop (cut off) of the servo pulse. The math is simple. Setting the cap to 1800us is done using this logic: If Pulse >1800us then Pulse = 1800us. The signal is not altered to compress or expand the pulse to fit within a range.
um, then regretfully it's not what matchboxes do then and isn't capable of matching. simply truncating the value means that a servo that needed its endpoint adjusted will be way out of sync until all ganged servos arrive at the end point. I don't know if I can even imagine a useful case for basic truncating...
Last edited by theKM; Oct 27, 2013 at 07:07 PM.
 Oct 27, 2013, 07:07 PM ↓↘→ + (punch) An example... ganging of two servos where things aren't ideally exacting, but perfectly possible with a retail servo matcher: one servo is running full from center of 1500 to 2000... second servo at the matched endpoint is at 1900... with truncating, the first servo will be at 1900 and the second servo that is meant to be working together is already at its truncated endpoint with 100us to go on the first servo. epa's need to be a multiplier so that when the input is at 1900us, the first servo is at 1900, and the second is at its ratio of 1820us. with truncating, there's literally no chance of the servos working properly together in the middle stages until the input is at its maximum, and at its worst possible match when the input reaches a truncation point. Last edited by theKM; Oct 27, 2013 at 07:13 PM.
Oct 27, 2013, 07:14 PM
↓↘→ + (punch)
Quote:
 Originally Posted by JimDrew I spent a few hours today experimenting with the new X10+ servo output code and RFU. I found that I can fly hundreds of feet up and away with the button pressed. In fact, I spent a little bit of time with no antenna on and the button pressed. I was able to taxi to the end of our runway (425 feet) from the middle (so about 200 feet) with no antenna and button pressed. The range of the RFU is incredible. I don't recommend you ever do this, and just because system with an RFU passes the normal range test without an antenna, please don't fly without the antenna! The new servo code worked perfectly, so I will be changing everything to that new code so we can have independent output reversing/centering/epas.
RFU epicness here we come!
 Oct 27, 2013, 07:17 PM Still Screamin'! So Aaron, does this mean your truncation blues are rectified?
Oct 27, 2013, 07:20 PM
↓↘→ + (punch)
Quote:
 Originally Posted by rvincent So Aaron, does this mean your truncation blues are rectified?
not at all
as mentioned, I can't think of a use case where truncation would be useful, can't match a wing with it (scenario in post #171 above). Not being able to do differential was a bit of a red flag I suppose, differential should be possible as output should still be proportional all the way to the end point.

...still happy about the coming of the RFU though.
 Oct 27, 2013, 07:25 PM ↓↘→ + (punch) I guess the wires on the requirements were crossed in how EPA is regarded in the Tx, a boundary that should never be crossed (which is good for all kinds of reasons). But in this scenario, it needs to be able to contract or expand the range of motion proportionally. the programming of it would be the multiplier starts at 1... take off the center from the input signal, multiply by the multiplier at as fine a precision as practical, make it relative to center again, and there's the output. values less than 1 will decrease the epa, values larger will increase. increasing will be subject to the input signal precision as to if it loses resolution, but on XPS's oversampled data it will still be buttery smooth the whole way, even greatly stretched outputs. Last edited by theKM; Oct 27, 2013 at 07:38 PM.
 Oct 27, 2013, 07:39 PM Thread OP I understand what you are saying. You are not talking about EPAs, which are hard cut off points for the throws. What you want is something that will give you a ratio from 1500us (center) to some amount of travel that is adjustable. So, you may want to compress 500us (1500us+500us) into 1850us of space. Likewise, you may want to expand 1850us to 2000us. There is some pretty complex math to do this, which add hundreds (potentially thousands) of microseconds of calculation time - BUT it would pre-calculated so it wouldn't affect the servo outputs. All of this code is fresh in my mind, so I will look into seeing what is involved in doing this. The automatching uses current to build a huge table, so no math is needed. We don't have that much RAM in the CPU being used with the X10+ but it does have some hardware features that expedite math, so it might be faster than I think. What servo are you trying to use that can use extended PWM to make it spin 180 degrees? If I have to go this route, I might as well test it with one of those. Any ideas on the user interface for controlling the ratio? It's no longer a microsecond value - it's more like a multiplier value that needs a very high resolution. That is the problem I see with the math - it's going to need a very large value to work with to allow for very small deviations. Last edited by JimDrew; Oct 27, 2013 at 07:47 PM.
Oct 27, 2013, 07:46 PM
↓↘→ + (punch)
Quote:
 Originally Posted by JimDrew I understand what you are saying. You are not talking about EPAs, which are hard cut off points for the throws. What you want is something that will give you a ratio from 1500us (center) to some amount of travel that is adjustable. So, you may want to compress 500us (1500us+500us) into 1850us of space. Likewise, you may want to expand 1850us to 2000us. There is some pretty complex math to do this, which add hundreds (potentially thousands) of microseconds of calculation time - BUT it would pre-calculated so it wouldn't affect the servo outputs.
this is why I was initially thinking the math was the reason for the output lag... the output is proportional up to the end point, even if the end point was less than or greater than the input value. This is how EPA's are in radios though, half travel on the stick represents half output of whatever the EPA was set to (some dual rates screens let you go 140%, which is a little silly, but it's still proportional to the EPA).

math, always gets in the way of world domination.
 Oct 27, 2013, 09:57 PM Thread OP I think I will keep the slider and microseconds for the user interface. I will just subtract the value from center and use that as the multiplier.
 Oct 28, 2013, 02:26 AM Registered User This sounds like individual rates for each servo and direction.
 Oct 28, 2013, 04:20 AM Registered User Hi there Hi Jim I watch forum here closely. There is always something new to learn here every hour around the clock. I do fly large scale gliders , tow planes and every set up must be carefully evaluated and checked all the time. Matching servos is crucial due to stiff linkage connection and powerful servos. Some times I need to use two servos on the same surface. Every servo is different regardless of the brand. Even with two identical servos you can spot differences how they move along the entire travel. I have never own the match box I have start to use servo stretcher from ServoCity to use on complicated kinematics of Landing Gears. Now I am using it in variety of application This device works with individual servo, but does what theKM is mentioning here. I am not sure how it deals with resolution issue (it seems like it is not great), but it makes me confident servos do not work against each other alongside the travel ZM Last edited by Smyk; Oct 28, 2013 at 04:37 AM.