ARUni, AR Universal board - Page 11 - RC Groups
Thread Tools
Feb 17, 2016, 04:49 PM
Registered User
Do you want to memorize the analog trim positions?
Programming a trim switch("Insta Trim") may be useful for the analog trimmers.
Otherwise, you can convert the analog trims to the digital trims if you're willing to put some labor
Sign up now
to remove ads between posts
Feb 17, 2016, 06:14 PM
S.A.D. member
ivanc's Avatar
Quote:
Originally Posted by renatoa
So was decided anything how to handle vintage analog trims, if even...
How about old school analog mixing? The values below provide approx. 30% mixing.



Ivan
Feb 18, 2016, 01:15 AM
Registered User
flygear's Avatar
I thinkthe board is almost getting there.

If all ok, we may be able to see real PCB in about a weekend.

1st pin of all header is assigned for 3.3V.
except I2C bus pin header "S", I don't feel safe if 5V is sitting next to the bus lines. 5V does damage MCU.

Feb 18, 2016, 01:28 AM
Registered User
ivan,
I don't think this ckt works well ....remember that the ADC input is high impedance. So the Joystick and Trim pots interact as parallel resistors each other, producing nonlinear voltage over the stick (or trim) movement. Your ckt would work if you use feed resistors of much higher values than the pots (which forms the source impedance) going into the virtual ground of an OP buffer before ADC. This gives you reasonably linear addition of the voltages from the two pots.

Well... this is the theory. If you look at some analog proportional systems, the trim is simply a "series" pot to the joystick, and no one seemed to have complained

So, the question is how much linearity you want for 10bit (11bit?) ADC. If you need only 5bits or so, your ckt as well as a series trim pot should work fine.
Personally, I still use a lot of analog servos, which seem only capable of stopping at around 10~20 points over the entire movement. This indicates that 5bit resolution is more than adequate for me

Another issue is calibration of the joystick center and the range. You need to set the trim at the right positions at each test.
I still use analog trims with ACE, but it is becoming less favorite in these days.
Feb 18, 2016, 04:56 AM
Registered User
flygear, (again) looks great and one vote for "let's go for it".

As for the analog trims, other than the mechanical trim which rotates the main pot extra angles to shift the neutral point,
a small pot (1.5K to 2K) is serially connected to the main pot (5K) and the neutral point moves linearly as the trim lever moves (simple Ohm's law).
The analog trims just work with AR9X's ADC without modifying anything - only catch is it doesn't memorize the trim position.
Feb 18, 2016, 09:23 AM
60 years of RC flying
Daedalus66's Avatar
It's not entirely clear what it would mean to memorize analog trims. On my 1996 JR XP783 you can transfer the trim settings to sub trim, then reset the trim levers to center. That's one way to do it and ErSky9x already includes that capability.
Feb 18, 2016, 09:40 AM
Registered User
flygear's Avatar
Alright, let's release it to the PCB maker.
Feb 18, 2016, 09:52 AM
Registered User
Great News! Thanks for the efforts, flygear

Daedalus66,
What flybabo was saying is that all the trim values are automatically stored at EACH Flight Mode. If you want to reset the trim lever BEFORE you change the FM during the flight, it is not only cumbersome, but dangerous...
Feb 18, 2016, 10:54 AM
S.A.D. member
ivanc's Avatar
Quote:
Originally Posted by signalone
ivan,
I don't think this ckt works well ....remember that the ADC input is high impedance. So the Joystick and Trim pots interact as parallel resistors each other, producing nonlinear voltage over the stick (or trim) movement. Your ckt would work if you use feed resistors of much higher values than the pots (which forms the source impedance) going into the virtual ground of an OP buffer before ADC. This gives you reasonably linear addition of the voltages from the two pots.

Well... this is the theory. If you look at some analog proportional systems, the trim is simply a "series" pot to the joystick, and no one seemed to have complained

So, the question is how much linearity you want for 10bit (11bit?) ADC. If you need only 5bits or so, your ckt as well as a series trim pot should work fine.
Personally, I still use a lot of analog servos, which seem only capable of stopping at around 10~20 points over the entire movement. This indicates that 5bit resolution is more than adequate for me

Another issue is calibration of the joystick center and the range. You need to set the trim at the right positions at each test.
I still use analog trims with ACE, but it is becoming less favorite in these days.
A trim pot in series with the control pot shifts the center of the control pot which I personally don't like. Also many of the old transmitters with separate trim pots use 5K or even 10K pots which will shift the center point of the control pot way off center losing too much resolution.

With the pots in parallel with summing resistors which simplifies wiring greatly as it can be done at the gimbal and you have only 4 wires for one gimbal going to the main board, you are getting a little non-linearity but you can take care of it in the firmware (ERSky9X or OpenTX) with expo. I've done it converting old transmitters with separate trim pots and it works great after I introduce a little expo for each gimbal set of pots.

BTW if the analog inputs have high enough impedance then you can use higher values summing resistors which will reduce the non-linearity - for example if the analog input impedance is 1M or higher you can use up to a 33K summing resistor for the control pot and up to a 100K for the trim pot. And of course you can add OpAmps between the pots and summing resistors - the low output impedance of the OpAmp will eliminate the non-linearity interaction of the control and trim pots through the summing resistors but they will also introduce noise and thermal drift.

Ivan
Feb 18, 2016, 11:00 AM
Registered User
Quote:
Originally Posted by Daedalus66
It's not entirely clear what it would mean to memorize analog trims. On my 1996 JR XP783 you can transfer the trim settings to sub trim, then reset the trim levers to center. That's one way to do it and ErSky9x already includes that capability.
How is this part "then reset the trim levers to center" done for a mechanical analog trim ?
Does this means that after transferring trims to subtrim, the analog trims values are ignored ? For how long... ? Until centered manually? Is there defined a deadband to detect centering ?
If not this way (analog trims ignored) how do you avoid to not apply trims twice: subtrim and actual trim ?

Quote:
Originally Posted by ivanc
A trim pot in series ...

With the pots in parallel with summing resistors ...
Ivan
You puzzle me with these connection schemas... aren't supposed the cursors of the sticks and trims are separate and connected to separate ADC inputs, in order to be able to read analog trim values ?!
Last edited by renatoa; Feb 18, 2016 at 11:05 AM.
Feb 18, 2016, 11:28 AM
S.A.D. member
ivanc's Avatar
Quote:
Originally Posted by renatoa
. . .

You puzzle me with these connection schemas... aren't supposed the cursors of the sticks and trims are separate and connected to separate ADC inputs, in order to be able to read analog trim values ?!
Maybe you are puzzled because you don't understand it. I offered a simple way to achieve what you described - using separate trim pots. Or maybe I don't understand what you are trying to do - happens all the time.

If you don't need S1 through S3 then you can wire three of the trim pots there and then assign these as trims in the firmware. One of the controls won't have a trim pot though in this case.

Ivan
Feb 18, 2016, 12:02 PM
Registered User
Well... I never flown with an analog trim but I have a very capable radio shell contemplating for this board to arrive.
The analog trims are pots, separate than the stick pots, not sure if this is obviously
In the original radio all the gimbal pots were connected to some op amps, each pot its own ope amp, not connected series or parallel before the op amps. Probably analog mixing of stick-trims is done by op amps, or maybe the op amps output separate values to the radio processor, no idea.

The question is how someone fly with such system ?

As asked above I understand the copy of trim values to subtrim, what I don't understand is what happens after... how is performed recentering, how to change/reset sub trims, etc...
Also, what happens if I am applying a trim in flight, than change flight phase? will the trim position influence the final channel value or not in the new flight phase... things like this
Feb 18, 2016, 01:08 PM
60 years of RC flying
Daedalus66's Avatar
Sorry if I confused things. I was simply describing how my old JR 783 managed to store analog trims for its eight model memories.

Just to be clear. After flying and setting the trim levers, you go into the menus and select the Trim Offset Memory function. This transfers the trim values to sub trim. You then return the trim levers to center and use Clear to exit. The result, if you do this consistently, is that trim levers for all models are centered.

That's all. I may be missing the point in the current context and will read more attentively.
Feb 18, 2016, 02:25 PM
RC Soarer & wheel re-inventer
debianhot's Avatar
Quote:
Originally Posted by flygear
Alright, let's release it to the PCB maker.
Great !! , one milestone less to cover...
Feb 18, 2016, 02:26 PM
RC Soarer & wheel re-inventer
debianhot's Avatar
Quote:
Originally Posted by Daedalus66
Sorry if I confused things. I was simply describing how my old JR 783 managed to store analog trims for its eight model memories.

Just to be clear. After flying and setting the trim levers, you go into the menus and select the Trim Offset Memory function. This transfers the trim values to sub trim. You then return the trim levers to center and use Clear to exit. The result, if you do this consistently, is that trim levers for all models are centered.

That's all. I may be missing the point in the current context and will read more attentively.
Same procedure for futaba/robbe


Thread Tools