×
Important information

To file a bug, find instructions here: Reporting bugs

Please only post images or files to this forum which you either own the copyright to or you are sure can be distributed under an open source license.

Devo7e work

More
15 Feb 2013 02:17 #6390 by PhracturedBlue
PhracturedBlue replied the topic: Devo7e work

skruth wrote: Another quick question about the replacement of devo7e's RF module with the one from 6s/8s.
Can anybody point me to a complete pinout of the 8s module (WK-21201RF). What I've digged from the devo8 connections and the flysky doc looks like:

the 8s module pinout (using the numbering in the flysky docs):
1: chip select
2: SCK
3: GND
4: MOSI
5: MISO
6: V+
10:Reset

The Devo7e:
1) Chip Select
2) SCK
3) MISO
4) MOSI
5) Reset
6) V+
7) GND

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 04:30 #6391 by vlad_vy
vlad_vy replied the topic: Devo7e work

PhracturedBlue wrote: the low pass filter will not cause the voltage to drop in that way unless there is a bug which there may well be.


Maybe MCU can be switched to 'sleep' mode in time ADC reading, for lowering noise?

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 05:02 #6393 by PhracturedBlue
PhracturedBlue replied the topic: Devo7e work
It turns out that the issue was that for some reason in continuous mode the ADC seems to pick up some coupling from the throttle channel. I've found I can still enable the low-pass filter without needing the ADC running fully in continuous mode. This fixes the voltage dependence on throttle without any negative effect. I've recalibrated the devo7e voltage and it should be withing 50mV from 4V to 5V.

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 05:34 - 15 Feb 2013 05:44 #6395 by sbstnp
sbstnp replied the topic: Devo7e work

PhracturedBlue wrote: It turns out that the issue was that for some reason in continuous mode the ADC seems to pick up some coupling from the throttle channel. I've found I can still enable the low-pass filter without needing the ADC running fully in continuous mode. This fixes the voltage dependence on throttle without any negative effect. I've recalibrated the devo7e voltage and it should be withing 50mV from 4V to 5V.


It looks this has fixed the voltage readout on Devo 10 as well (at first look). I'll keep and eye on it and update the issue on BB.

EDIT: I spoke too soon, I got 10.05V reading and the dreaded "Battery too low can't save" at the same time lol. I'll update issue #234

Devo 10 + Deviation = LOVE
Last Edit: 15 Feb 2013 05:44 by sbstnp.

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 08:18 #6398 by skruth
skruth replied the topic: Devo7e work
Great, thanks a lot.

Cheers, Stefan

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 13:59 #6404 by vlad_vy
vlad_vy replied the topic: Devo7e work

sbstnp wrote: EDIT: I spoke too soon, I got 10.05V reading and the dreaded "Battery too low can't save" at the same time lol. I'll update issue #234


Also, Devo 8, battery reading 4.77V and "Battery too low can't save" at the same time. Battery alarm 4.30V, batt_critical=3800.

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 14:07 #6406 by sbstnp
sbstnp replied the topic: Devo7e work

vlad_vy wrote:

sbstnp wrote: EDIT: I spoke too soon, I got 10.05V reading and the dreaded "Battery too low can't save" at the same time lol. I'll update issue #234


Also, Devo 8, battery reading 4.77V and "Battery too low can't save" at the same time. Battery alarm 4.30V, batt_critical=3800.


Mine was 7.5V critical, 8.8V alarm.

Devo 10 + Deviation = LOVE

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 14:31 - 15 Feb 2013 14:32 #6407 by Xermalk
Xermalk replied the topic: Devo7e work
Battery reading is correct on the 7E, however you always get a "Battery too low can't save" no matter the battery level.
You even get it with the charger connected.

Battery alarm 4.40V, batt_critical=3800

Im going to have to downgrade to a earlier version so i can fly tonight.

Its one helluva bug though, to bad you really cant increase or decrease the battery charge level simply by moving a stick :lol:

Does anyone know what the last working version was for the 7E?
Last Edit: 15 Feb 2013 14:32 by Xermalk.

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 16:22 #6408 by PhracturedBlue
PhracturedBlue replied the topic: Devo7e work
I can't really explain why the changes I made would have this affect. I also can't reproduce it on my 7e Tx (I think I get a battery alarm beep, but there are so many beeps now, I can't be sure what is actually beeping. I haven't had time to check any others yet.

I checked in a change that should prevent reading a glitch value which may prevent the battery low issue. Or not. Can't test until this weekend.

If a couple folks could test the latest code and see whether it actually helps of not, that would be helpful.

Please Log in or Create an account to join the conversation.

More
15 Feb 2013 16:47 - 15 Feb 2013 17:10 #6410 by sbstnp
sbstnp replied the topic: Devo7e work

PhracturedBlue wrote: If a couple folks could test the latest code and see whether it actually helps of not, that would be helpful.


The warning hasn't popped up in the last 2 minutes or so since I flashed my radio. I'll keep an eye on it for the next 15 minutes or so.

It used to appear on the 1 minute mark where I have set my low voltage alarm.

EDIT: ~20 minutes, no warning so far. Voltage looks good I lost around 0.2V with brightness set on 1.

Devo 10 + Deviation = LOVE
Last Edit: 15 Feb 2013 17:10 by sbstnp.

Please Log in or Create an account to join the conversation.

  • rbe2012
  • rbe2012's Avatar
  • Offline
  • So much to do, so little time...
More
15 Feb 2013 19:23 #6414 by rbe2012
rbe2012 replied the topic: Devo7e work
I have a similar problem on Devo8s - moving aileron to the left (mode2, right stick) let the voltage display 5,47V, in the middle 5,07V and to the right 4,83V (4xNiMH).
I have never seen this bevor but I am not watching the battery voltage when flying...
Version: devo8-b79f385a1569
While charging: 6,02 - 5,63 - 5,33
Tomorrow I can flash an actual version.

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 06:35 #6426 by vlad_vy
vlad_vy replied the topic: Devo7e work
Devo 8. Now I have not "Battery too low, can't save", but sometimes I see accidental throws 10-40% at stick channels (Input monitor, Channel monitor and Main screen). 7e058aa9d523

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 06:58 #6427 by cmpang
cmpang replied the topic: Devo7e work
went out to the boating pool today to perform field tested my 7e after Tx module shorting diode mod along side with my Devo 10. Confirm that radio signal is as good (if not better) as the 10.

method is to lower the power output bit by bit until the Rx just goes into failsafe mode. The critical settings for both Txs turned out to be the same value of 1mW.

Be reminded that both antennea have to be pointing in the same orientation during the test, as the original 7e antenna is unable to fold up during use and would natually lay parallel to the ground. While in the Devo 10 we usually points the antenna up to give the best radiation pattern.

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 10:53 #6428 by PhracturedBlue
PhracturedBlue replied the topic: Devo7e work

vlad_vy wrote: Devo 8. Now I have not "Battery too low, can't save", but sometimes I see accidental throws 10-40% at stick channels (Input monitor, Channel monitor and Main screen). 7e058aa9d523

okay. something is just broken here. I'll revisit the whole ADC thing again. Maybe for all but the devo12 go back to the old way of reading which seemed to work fine on all other tx.

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 11:13 - 16 Feb 2013 12:10 #6430 by Xermalk
Xermalk replied the topic: Devo7e work
Something is very broken, there seems to be random input on all transmitter channels.

Edit: Theres a very high chance that one of my servos have been reversed from the random input. The joy of alpha firmware :P

Edit 2: called wolf to early. Seems i somehow corrupted my model.ini, so servo isnt reversed.

deviation-devo7e-d7e61bcd6c44.zip is still sending garbage over the channels though. And the channel monitor screen flickers when it does. Reverted to b76ac7445568 for now.

Edit 3: Voltage measurement seems to be working just fine in d7e61bcd6c44, PB you might want to hunt down exactly in what revision it broke.

File Attachment:

File Name: deviation-...6c44.zip
File Size:316 KB
Attachments:
Last Edit: 16 Feb 2013 12:10 by Xermalk.

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 13:53 #6433 by vlad_vy
vlad_vy replied the topic: Devo7e work
Anduino example of noise reduction with use 'sleep' mode
arduino.cc/forum/index.php/topic,38119.0.html
int rawAnalogReadWithSleep( void )
{
  // Generate an interrupt when the conversion is finished
  ADCSRA |= _BV( ADIE );

  // Enable Noise Reduction Sleep Mode
  set_sleep_mode( SLEEP_MODE_ADC );
  sleep_enable();

  // Any interrupt will wake the processor including the millis interrupt so we have to...
  // Loop until the conversion is finished
  do
  {
    // The following line of code is only important on the second pass.  For the first pass it has no effect.
    // Ensure interrupts are enabled before sleeping
    sei();
    // Sleep (MUST be called immediately after sei)
    sleep_cpu();
    // Checking the conversion status has to be done with interrupts disabled to avoid a race condition
    // Disable interrupts so the while below is performed without interruption
    cli();
  }
  // Conversion finished?  If not, loop.
  while( ( (ADCSRA & (1<<ADSC)) != 0 ) );

  // No more sleeping
  sleep_disable();
  // Enable interrupts
  sei();

  // The Arduino core does not expect an interrupt when a conversion completes so turn interrupts off
  ADCSRA &= ~ _BV( ADIE );

  // Return the conversion result
  return( ADC );
}

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 14:26 #6435 by PhracturedBlue
PhracturedBlue replied the topic: Devo7e work
vlad, the model you are implying will not work for us. We need to be doing work all the time (handling limers, sending data to the transceiver. we cannot shutdown while reading the ADC. This problem isn't due to noise anyway. This current problem is due to shutting down the adc in the middle of a conversion.

What we now know:
On the devo12, the board wiring is introducing significant noise. We need low pass filtering here otherwise there is a lot of jitter

Continuous-read works well on the devo12, but seems to affect the voltage-sense value for an unknown reason.

shutting down the adc via an interrupt can cause a partial-write which appears as a channel glitch.

I will run further experiments. Ideally I'd like to use the same methods on all transmitters. But if I can't make that work, I'll use the old single-sample solution on the devo6/7e/8/10 and the low-pass solution on the devo12.

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 18:48 #6444 by PhracturedBlue
PhracturedBlue replied the topic: Devo7e work
Ok, after a lot more testing, I have checked in a solution which seems to give stable voltage readings on all Tx, and does not exhibit the glitch behavior.
Let me know if you see any more issues.

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 18:51 #6445 by sbstnp
sbstnp replied the topic: Devo7e work

PhracturedBlue wrote: Ok, after a lot more testing, I have checked in a solution which seems to give stable voltage readings on all Tx, and does not exhibit the glitch behavior.
Let me know if you see any more issues.


Flashing it right now. Will test tomorrow since it's almost 9PM here.

Devo 10 + Deviation = LOVE

Please Log in or Create an account to join the conversation.

More
16 Feb 2013 18:55 #6446 by FDR
FDR replied the topic: Devo7e work
You put the battery voltage measurement into the stick readout routine, didn't you?
I just wanted to suggest something like that to prevent the need for 2 kind of ADC reading... :)

Please Log in or Create an account to join the conversation.

Time to create page: 0.359 seconds
Powered by Kunena Forum