AttoPilot with video transmitter (FPV) - Interference and Tuning questions

[vassilis]

I had my AttoPilot for a few months now and I have been very happy with the whole package, however in my latest setup I seem to be having some issues with interference from my video transmitter and some problems getting the roll tuning right (may be related).

First about my setup:
AttoPilot v1.8
EasyStar with aileron modification
900MHz / 500mW video transmitter
Custom OSD and telemetry system (using the AttoPilot telemetry stream)

I noticed the interference on the first flight, first the battery current reading was noticeably off, then on the next flight the airspeed was off by about 50kph. I then noticed that the offsets were present only when the transmitter was close to the AttoPilot. All my servo wires are twisted and as per the manual I make sure that I first power the AttoPilot, set home position and then power the video transmitter. Also I should note that I have no other interference issues on the plane, GPS is reading more than 7 sats during the flight and there is no interference with the servos or receiver. I also tried wrapping the AttoPilot in foil but that did not help either.

Any experience out there with AttoPilot and video transmitters? If you had similar issues what did you do?

I will make a separate post about my tuning problems.

http://www.youtube.com/watch?v=jsPFhByGdaI

Thanks!
Vassilis

[Irv]

Vassilis,

A very clean looking installation.

I flew my ES with a 900Mhz video system configured much like yours, but without an AttoPilot.

I had problems with 900Mhz interference getting into the pan and tilt servos.
I wrapped several turns of each of the servo wires around toroids. That cleared the jitter up.

My AttoPilot installation in an Easy Glider had problems until I moved the XY sensor away from the motor/ESC wires. I also wrapped the throttle wires through a toroid. That improved the performance a lot.

Hope this helps.

Irv

[2jers]

Kalimera Vassili.

Also, if you have "long" cables, you can use clamp-on ferrite cores, like this one:http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=240-2124-ND


Very nice OSD, and looks very functional

[dmgoedde]

I saw this with a customer here in Arizona when the 1 Watt 9XTend was turned on before AttoPilot. The interference might be through-air or being picked up via the wires to the analog sensors (XY and Z thermopiles, power sensor). It seems that the 8 channel ADC gets confused, thus pitch, roll, airspeed, and power sensing will all be off. Curiously if Atto is turned on first and THEN the RF transmitter, the issue went away in two specific installations that I refer to. So, it seems to be something with initialization of the ADC. In all cases I saw, the processor and data logging, Rx reading and servo control were fine (and altitude barometric pressure was unaffected, which is an SPI chip and not one of the analog channels). What I am saying here is only the ADC got affected.

Try turning down the transmitt power if you can, moving the RF sources further from Atto and leads going to Atto, putting ferrite toroids on the XY, Z, and power sense leads, and also turn Atto on before the RF power is applied.

[vassilis]

Thanks all for the comments!

@Irv, 2jers
I have a toroid on the ESC wire. I also already have a few of the clamp-on toroids and I will try them out. However I do not think the pickup is from the wires as I am able to get rid of the offset by moving the transmitter further away. It must be picked up through the air.

@dean
Moving the transmitter in another location will be a major re-work of the model as I already have some of the wires in the fuselage. Also as you can see there is not much free space!

I always power Atto and set home position before I power the video transmitter. Unfortunately I am not able to reduce the transmit power.

I do notice problems with the roll during flight, but the main problem is airspeed. There is currently no way to fly the model in ARC or Auto mode since the airspeed can vary up to 40kph causing the motor to shutdown and model to stall at times.

My plan is try one of those: http://www.dpcav.com/xcart/product.php?productid=16417&cat=270&page=1

Vassilis

[Gary Mortimer]

Is it all sharing the same power source??

Any noisy bit will shine through that way, common earth and all.

[vassilis]

@gary

That is a good point, I do have a separate power source for the camera and video transmitter however they share the same ground as the main battery because of the OSD (requires a connection to the telemetry port of the AttoPilot and the camera/transmitter). However, I still think that this is over the air interference (since moving the camera POD away from Atto seems to have an effect). I will be trying a flight without the OSD this weekend (weather permitting!).

@2jers

Regarding the OSD, I will be releasing the schematics and source code as soon as I get the whole setup working. I also send the Atto telemetry stream through the audio channel of the video transmitter.

[2jers]

@2jers

Regarding the OSD, I will be releasing the schematics and source code as soon as I get the whole setup working. I also send the Atto telemetry stream through the audio channel of the video transmitter.


Thank you

[airmcn_3]

Regarding the OSD, I will be releasing the schematics and source code as soon as I get the whole setup working. I also send the Atto telemetry stream through the audio channel of the video transmitter.


Please do that!

[vassilis]

Quick update, weather was too bad today to fly but I completely separated the grounds of the video transmitter and the Atto and the interference is still there. So to sum up, separate power sources including grounds and the interference persists. Moving the transmitter further away does help.

The 900MHz antenna filter should be arriving early next week, if that does not help then I will have no choice but to move Atto to a bigger model.

Vassilis

[airmcn_3]

Quick update, weather was too bad today to fly but I completely separated the grounds of the video transmitter and the Atto and the interference is still there. So to sum up, separate power sources including grounds and the interference persists. Moving the transmitter further away does help.

The 900MHz antenna filter should be arriving early next week, if that does not help then I will have no choice but to move Atto to a bigger model.

Vassilis


This is very unusual. I have had 1W of 900MHz sitting darn near on top of Atto. On the other hand it was a modem and frequency hopping at that. I had a similar issue with 2.4GHz video at 1W but it affected GPS not anything else. The key was distance....

[vassilis]

I think the difference between a data modem and video transmitter is bandwidth, the video signal is much more broadband. Also I am pretty sure that the signal of your data modem is much cleaner than the video transmitter.

I do agree about the distance! I still have some hope for the low pass filter...

[vassilis]

Unfortunately the antenna low pass filter did not help much. I am left with no option but to move the AttoPilot to a larger model (Telemaster 6). In this model the video tx is about 24 inches behind the electronics. Hopefully that will solve my problems!

Following is a video from the last flight, you can see the large roll offset, when I switched to ARC the model turned violently and almost dived to the ground.

http://www.youtube.com/watch?v=vDj_2fxuBtI&fmt=18

Vassilis

[airmcn_3]

Unfortunately the antenna low pass filter did not help much. I am left with no option but to move the AttoPilot to a larger model (Telemaster 6). In this model the video tx is about 24 inches behind the electronics. Hopefully that will solve my problems!

Following is a video from the last flight, you can see the large roll offset, when I switched to ARC the model turned violently and almost dived to the ground.

http://www.youtube.com/watch?v=vDj_2fxuBtI&fmt=18

Vassilis


Moving up to a larger model is not the end all be all, I use Atto in a 1.3m flying wing with no issues, instead of moving to a new aircraft why don’t you change the video frequency? I can get good quality video at 3.5km without an issue on 5.8GHz.....

Are you sure some of the reading is not from an improper thermopile installation? Sorry if this seems harsh but the XY head is not in a good spot and if is blocked just right will give you bad readings.....

You should not have had the aircraft dive towards the ground when going into ARC, that is a stabilization mode and only uses inputs by the pilot and simply stabilizes it. If there is an issue with your XY head this will cause you to go haywire in ARC........ There is no reason for a dive other then a bad or blocked XY head.

Also be darn sure you have the sticks centered when going into ARC, if you have down elevator that is what the aircraft will think is level initially, it takes Atto a few seconds to recognize and self trim the aircraft.
Chris

[vassilis]

@airmcn

Your comments are not harsh, any advise on my setup are welcome!

Regarding the 5.8 ghz tx, it is an option however I would like to see how sensitive Atto is at 900mhz, increased spacing should do the trick.

About the xy head installation, I used to have a CPD4 xy head installed at exactly the same spot. I had no problems with stabilization. In my larger model I have a different installation that may help.

Could it be a hardware problem with my xy or z sensors? They do respond to roll and pitch, anything else I can check?

I am certain the sticks were centered.


Vassilis

[airmcn_3]

@airmcn

Your comments are not harsh, any advise on my setup are welcome!

Regarding the 5.8 ghz tx, it is an option however I would like to see how sensitive Atto is at 900mhz, increased spacing should do the trick.

About the xy head installation, I used to have a CPD4 xy head installed at exactly the same spot. I had no problems with stabilization. In my larger model I have a different installation that may help.

Could it be a hardware problem with my xy or z sensors? They do respond to roll and pitch, anything else I can check?

I am certain the sticks were centered.


Vassilis

Although chances are that your xy head is bad are slim but it is an electronic part and anything is possible.

The only way to check it is with a special firmware written specifically for the purpose of checking X,Y and Z sensors. It requires you to flash the firmware, open the propeller prop plug from Parallax and then you would tell it to load ram instead of firmware, this flashes the latest firmware you have selected for this test only. As soon as you un-plug your prop plug from the bird it will default back to the firmware you were previously using.

I will talk with Dean and see if I can pass along the file. Do you have a parallax prop plug for burning firmware?

Chris

[vassilis]

Chris,

I have a prop plug and have aleady updated the firmware. The "sensor test" firmware sounds like a good idea, please let me know if I can try it.

Thanks!
Vassilis

[dmgoedde]

@airmcn

Your comments are not harsh, any advise on my setup are welcome!

Regarding the 5.8 ghz tx, it is an option however I would like to see how sensitive Atto is at 900mhz, increased spacing should do the trick.

About the xy head installation, I used to have a CPD4 xy head installed at exactly the same spot. I had no problems with stabilization. In my larger model I have a different installation that may help.

Could it be a hardware problem with my xy or z sensors? They do respond to roll and pitch, anything else I can check?

I am certain the sticks were centered.


VassilisHi Vassilis - If XY head sticker number facing forward is correct in SET.txt, and your sticks are centered while transitioning from Manual --> ARC (so that nuetral is learned properly) there are still 2 things that can cause stability to go haywire:
1) Improperly tuned pitch or roll gains. Mostly if roll gain is grossly too low, then the model will readily roll over to a tipstall like condition. I assume you have already done good flights in ARC and Auto without video RF turned on though, and the system is well tuned.
2) RF power will corrupt the analog signals coming into Atto, and/or the ADC doesn't boot up properly. You can test this safely on the ground during startup sequence when GPS first achieves lock and it is time to set the home position by going to Auto mode briefly. This initial Auto mode is a 'freebie' in terms of the 5Hz navigation loop has not yet started in the code and the motor PID control loop is in a pauese, and the motor will not come on. So when you have video RF power on, get GPS locked, then switch to Auto and stay there for a minute, and hold the plane up and bank it around to verify the ailerons and elevator move in the direction that would return the model to level flight. If problems exist from the RF energy, it should be apparent on the groud in this test. You can repeat the entire test (power down Atto and restart) but this time with video Tx off.

I have routinely used the 900MHz 9XTend modem at full 1000mW power, but if too close to Atto see issues like Atto not booting up properly, or the servos 'talking' in sync with the telemetry output strings.

[vassilis]

Just got back from a test flight on my Telemaster, I would call it a mixed success.

The video tx is all the way to the tail, next to the stabilizer.
The spacing from the video tx and the Atto is 28 inches
The spacing from the video tx and the xy head is 21 inches

No interference in the airspeed was noticed (I assume 10-20 kph at ground is a normal offset since I have not seen it lower even with the tx off).

However, during the flight the roll offset in ARC mode is still there!!! Here is what happened step by step:

- I wait for gps lock (video tx off)
- Once lock is achieved I set home position (video tx off) by switching to Auto mode with the right stick centered.
- In Auto mode with the model on the ground, I see the ailerons maxed out for a left turn (video tx off, roll and pitch gain set to 40).
- While still in Auto mode, to move the ailerons to neutral I have to bank the model by about 20-30 degrees (video tx off)
- The elevator seemed to be fine with neutral archived at close to 0 degrees pitch (video tx off).
- I switched back to manual turned the video tx on and launched the model.
- When I switched to ARC mode (with the right stick centered), the model would not hold zero roll even after waiting several seconds for the bias correction to take place.

Let me know if you would like to have a look at the log file.

At this time, I do not believe that I am looking at an interference problem. Either my expectations of ARC mode are too high (in terms of holding roll) or there is a problem with my hardware. I should note this is the exact same behavior I get with my EasyStar in ARC mode.

Also, on my EasyStar I used to have a CPD4 co-pilot and it would hold roll and pitch remarkably well with very little change in heading with the right stick at neutral. I expected that the AttoPilot will be able to do at least as well.

Regarding PID parameter tuning, on the EasyStar I have tried about 7 different gain values, ranging from 15 all the way 75. I have never been able to achieve good roll control (pitch has been fairly good however).

I am sure there must be something wrong my setup since other people had success with Atto, I am just not sure where to look next. Any way I can try the test firmware to verify that the sensors are operating correctly?

Thanks,
Vassilis

[vassilis]

Attached is the log from my last flight, I moved Atto back to my EasyStar and flew with the antenna filter and a different video transmitter (still at 900MHz).

This time, since I am running out of things to try, I got the roll gain all the way up to 90! It did not make much difference in ARC mode, however I did have oscillations in Auto mode.

From the log it can be seen that in ARC mode and zero commanded roll the plane goes in circles even though the roll shows as zero! Simply plot the PID and look at the heading instrument when the roll setpoint is zero degrees.

Vassilis

[Gary Mortimer]

Why dont you remove the video all together and get your airframe flying well and then introduce other elements?

Good luck

G

[vassilis]

@Gary

I have, my very first flight when I first got Atto was without the video tx. At the time I blamed the roll problem on improper gain tuning. I will try again with the latest setup, but I am not convinced the roll problem is interference related (as in the case of the airspeed). On my Telemaster the long separation of Atto and the video tx did fix my interference problems, however the roll problem persisted.

@Dean

As you recommended, I tested the roll response before launch when setting home position. To confuse things even more: when holding the EasyStar over my head, with and without the video transmitter I noticed good response on the ailerons and elevator when moving the model. At about zero degrees roll and pitch the ailerons and elevator where at neutral even when the video transmitter was on. I did not get the same behavior on my Telemaster, but it may be because the model was on the ground and I was standing relatively close.

Vassilis

[vassilis]

As advised I tried a flight without the camera, video transmitter and OSD to keep things simple and establish if the roll problem is EMI related.

The roll problem persists even without the video transmitter!

The attached log file has two flights, first flight with video transmitter and second without. I used the same SET file for both. As you can see, I get the same behaviour on both flights. Atto things roll is zero but in fact the model is banking and turning fairly fast (see the heading while the commanded roll is zero). When flying with the video off, in one occasion, I had to save it from a dive. I know the PID tuning can be improved, however I have tried gains from 10 all the way to 90 with the same problem.

Sticks are centered before switching to ARC mode (to read the neutral possition) and parameter $45 is set correctly. Ground testing when setting home position seems fine.

So far I can tell the following:
- Roll problem is not related to video interference. The test today proves that. In addition I had the same problem in two different models so even though I do not rule out something wrong with my setup, it is unlikely.
- Airspeed and mAh offsets: These are EMI related, however I can leave with it, the airspeed error during flight is small.

Any further advise will be appreciated! Right now I suspect the firmware, XY/Z sensors or Atto module itself. By the way I am using firmware version 4_14_09, what is everybody else using?

Vassilis

[airmcn_3]

As advised I tried a flight without the camera, video transmitter and OSD to keep things simple and establish if the roll problem is EMI related.

The roll problem persists even without the video transmitter!

The attached log file has two flights, first flight with video transmitter and second without. I used the same SET file for both. As you can see, I get the same behaviour on both flights. Atto things roll is zero but in fact the model is banking and turning fairly fast (see the heading while the commanded roll is zero). When flying with the video off, in one occasion, I had to save it from a dive. I know the PID tuning can be improved, however I have tried gains from 10 all the way to 90 with the same problem.

Sticks are centered before switching to ARC mode (to read the neutral possition) and parameter $45 is set correctly. Ground testing when setting home position seems fine.

So far I can tell the following:
- Roll problem is not related to video interference. The test today proves that. In addition I had the same problem in two different models so even though I do not rule out something wrong with my setup, it is unlikely.
- Airspeed and mAh offsets: These are EMI related, however I can leave with it, the airspeed error during flight is small.

Any further advise will be appreciated! Right now I suspect the firmware, XY/Z sensors or Atto module itself. By the way I am using firmware version 4_14_09, what is everybody else using?

Vassilis


4_14_09 is the correct firmware.

I just looked at your SET file and WOW you have a lot of tuning to go.

I am going to give Dean a call and have a chat with him, we will try and determine if it’s your XY head or not.

Looking at your SET file I see many things that are way off. As you may or may not know we have flown V1.8 on all types of airframes, never have I see the numbers you have in your SET file. If I was to input your set file into any aircraft we have flow it would crash. NOW I am not trying to bash you by any means I really want to help you get this figured out. I suspect the reason for the wild numbers in the SET file is due to bad hardware.

Your LOG file shows some weird plots. If I was to guess I would say it’s the XY head.

I will get back to you ASAP.

Chris

[vassilis]

Chris,

I appreciate the quick reply and as you can tell I can use some help!

Regarding the SET file, I simply followed the instructions in the manual. I have left most values to defaults (the SET file that came with Atto) and modified the ones as per the manual. If you are referring to the PID gains, I was also surprised that I could not get the roll to oscillate even with a gain of 90.

Vassilis

[saabguyspg]

Vassilis can you post some info on your wonderful osd and telemetry? possibly some sort of diagram? I assume you are getting telemetry from atto down the audio line? That is great because it would save me adding another device to the plane!

Have you been able to solve your problems?

[vassilis]

As I promised I will release the design (schematics and source code) soon (I am hopping this weekend). Please keep in mind this is a DIY project and work in progress (i.e. incomplete), if you choose to re-use be prepared to do some programming on your own.

I will make an other post with a link to the files as soon as they are ready!

Vassilis

[saabguyspg]

Great, sorry I am not trying to be pushy just curious and very impressed all at the same time! Good luck and I look foreward to what you create!

Thanks again.

[vassilis]

You can find the design files of the OSD in my blog:

http://vrhome.net/vassilis

Let me know if you find it useful!

Vassilis

[airmcn_3]

You can find the design files of the OSD in my blog:

http://vrhome.net/vassilis

Let me know if you find it useful!

Vassilis


Very nice! Thanks for that.

Chris