Help?

[bernieL0max]

Hey there,
I've had the parts sitting around for months to built a Tri, including a Paris v3.0 board, Arduino Pro Mini 328, etc.

I assembled the board a couple of nights ago, originally with boards from second-hand WM+ and NK, I'm a tech, so no difficulties soldering up the boards, etc. I uploaded MultiWii v1.9 with only a few basic changes to the config.h for my RC Timer 30A Speed Controllers - which aren't even connected at this stage.

If I power up the board, connect the FTDI, and load MultiWiiConf, it connects to the device, and displays the control inputs from my JR R770S without any problem; what it does not show is any values from the accellerometers or gyros.

I convinced myself that the second-hand WM+ and/or NK may have been faulty, so today I went and brought brand-new ones. At this stage I have only connected the WM+, yet I am still not seeing any readings from this in MultiWiiConfig.

Any thoughts or advice? Any more info you need to be able to assist? I'm trying to find my digital camera to take some shots of the boards.

Thanks in advance!

Matt

[UndCon]

You will not see the movement of the Tricopter figure in bottom right corner until you solder in the NK
But you should see graphs from the WM+
Assuming you soldered the senor correctly and you connected to the board correctly in GUI - what status do you have on the internal pullups?



On my Paris V4 i have:

//enable internal I2C pull ups
#define INTERNAL_I2C_PULLUPS

//UndCon

[bernieL0max]

yes, INTERNAL_I2C_PULLUPS is enabled.

The only lines I had changed from default were MINTHROTTLE to suit my ESCs, and #define TRI of course
Everything else was default, although I've turned my attention away from the sensors to play with other stuff for a minute, and have enabled #define SERVO_TILT... the pan/tilt servos centre, but that's about it

I've just hooked up the speed controllers and yaw servo... the yaw servo is functioning as expected. the speed controllers themselves are 'armed' on bootup (they beep constantly if they're not), and if I force it to ARM by setting the all conditions of AUX1 and AUX2 to "ARM"; the motors spin, and throttle/aileron/rudder/elevator stick inputs effects the relative motor speed as expected.

I find it hard to believe I've managed to get 2 dud WM+(?!). Will take some pics shortly and post.


I read an obscure post about needing to disconnect the FTDI 5V pin to get readings from the sensors... I have just slapped together a special cable that allows me to switch the 5V line off and on; this did not improve the problem. Am i misunderstanding?, does it mean I should cut the actual USB +5V RAIL(?!)

[mon_lolo_fr]

I read an obscure post about needing to disconnect the FTDI 5V pin to get readings from the sensors... I have just slapped together a special cable that allows me to switch the 5V line off and on; this did not improve the problem. Am i misunderstanding?, does it mean I should cut the actual USB +5V RAIL(?!)

You must disconnect the +5V on the FTDI end only if one of your BEC is supplying something on your board (Arduino, WMP, Radio RX, etc...).

What I'm doing to check the GUI, I connected the +5V on the FTDI end and supply also the RX with it, BUT I DISCONNECTED ALL ESCs BEC from the board (I check my ESCs independently and configure them all at a time) .

[bernieL0max]

I have a BEC connected to the "BEC" header on the board, and have tested that it is powering the board/RX/etc, all the 5V rails are connected together, so this is sufficient to power my RX without splitting the BEC or running a second one. Have just taken some pics, now to try and shrink them down to a size suitable to upload...

[bernieL0max]

Pics...

WM+ wiring

WM+ connected to board

[bernieL0max]

And a shot of the board, BEC, FTDI; you can see the switch I put on the 5V line, so I can undo the mod with the flick of a switch previously the FTDI board was soldered onto the Paris board, but I needed to make it removeable anyway in order to fit the whole thing into a camera dome.

wider shot

[bernieL0max]

Thanks for the input so far guys, very much appreciated. It's almost 2am here, I'd better hit the sack. Hopefully someone will help me work out what's going on

[UndCon]

I see a problem - move VCC to 3V3

[mon_lolo_fr]

I see a problem - move VCC to 3V3

The WMP has its own onboard voltage regulator, why moving VCC to 3.3V ?

[bernieL0max]

OT: for a moment, I have to admit, I find it more than a ltitle bit frustrating that I have to rely on google searches to find photos from various websites and forum posts to try to discover the actual wiring for the Paris boards, and with the release of v4.0, multiwii.com have removed all mention of the v3.0 boards from their site(?!). I find photos showing completely different orientations of the WM+ and NK boards, VCC hooked up to 4V5 and 3V3, I google and scoure forums, and read threads about all sorts of things before I find that it has its own internal regulator.

Shame we cant have 1 [set of?] factual/instructional webpage(s) or document(s) that clearly and concisely communicate all of the important/essential information and steps.


Back On Topic... still no luck at my end. I've gone over the board with a magnifying glass and multimeter, no shorts/brigdes anywhere, and it is a brand-new WM+, so surely I can rule out a faulty board? Should I maybe try and connect it to the other I2C bus?, will this simply work without any additional configuration?

[mr.rc-cam]

I have to admit, I find it more than a ltitle bit frustrating that I have to rely on google searches to find photos from various websites and forum posts to try to discover the actual wiring for the Paris boards, and with the release of v4.0, multiwii.com have removed all mention of the v3.0 boards from their site(?!).

I have a Paris V3 and I will admit that formal documentation was sparse (just a few example photos are all I ever found). But now that they have removed all V3 material from their site I can see how it would be a whole new level of frustration.

Back On Topic... still no luck at my end. I've gone over the board with a magnifying glass and multimeter, no shorts/brigdes anywhere, and it is a brand-new WM+, so surely I can rule out a faulty board?

For now, let's assume your WMP is ok and you don't have any wiring issues. If you closely follow these ten steps I expect you will have better success with your TriCopter:
1. Download MWC V1.9 again. Yes, download it again. Verify it has the Nov-12-2011 date stamp. Use this file set in place of what you have now, even if it appears to be the same.
2. Edit the V1.9 config.h and enable TRI (uncomment #define TRI). Do NOT enabled the internal pullups. Avoid any other edits for now (you can re-edit config.h after the WMP sensor is working).
3. Verify that your Arduino Pro Mini 328 is the 5V version and not the identical looking 3.3V CPU. Do this by measuring the voltage on the Arduino's Vcc pin. If 5.0V, you are good to go. If 3.3V, then I recommend you get the 5V CPU version.
4. On the Paris V3 board, disable the 2.2K pullup option on the bottom of the shield (remove solder from the solder bridge pads). Do not use these pullups!
5. Wire the WMP as before, but for the new V1.9 firmware please connect the WMP's Vcc power pad to 5.0V instead of 4.5V (D12) or 3.3V.
6. Add external 2.2K pullup resistors to SDA and SCL. The resistors must be sourced with 5.0V (not 3.3V!).
7. Power it all up and verify that the WMP has 5.0VDC at its power/gnd pads (measure at the WMP board, not elsewhere).
8. Launch the GUI and press start. Verify that the R/C controls are working by watching the GUI.
9. Watch the Debug2 status window. If the value is greater than zero then your sensor is bad, wiring is wrong, or some other hardware issue needs attention.
10. If Debug2 remains zero and your gyro data changes as you move the model, then pop the cork on the bubbly. Otherwise, report the debug2 value and all symptoms. Plus post some new photos that may help us see what you've wired up.

[bill516]

How about removing the Arduino and WMP from the Paris board and connecting them together as stand alone units as in the connecting elements page, you can rule out paris board that way.

Might have this wrong but I'm sure I got no GUI model movement with just a WMP connected to my Arduino it was only when I got a Nunchuck connected that I saw movement of the GUI model although the gyro data readings did move when I moved the actual model.

[UndCon]

That is true - there is no movement of your *copter in GUI before you install NK - but the curves work either way...

//UndCon

[bernieL0max]

Thanks Mr.RcCam, I will try the things from that, that I havent already. Little bit disappointed to spend money on a Paris board and then have to add a bunch of extra wires and components (beyond the WM+, NK and standard wiring of course!).

I will skip forward though, and wire up the NK, as at the moment I only have the WM+... I was under the impression the WM+ on its own was enough to get basic attitude readings, and the NK was required for autolevel functionality(?).

[mr.rc-cam]

I will try the things from that, that I havent already.

Approach all 10 things on the list as if you haven't done them yet. Start fresh, don't rely on the results from prior troubleshooting.

I will skip forward though, and wire up the NK, as at the moment I only have the WM+.

I recommend you leave the NK out of the circuity until you have zero I2C errors (debug2) and valid/reliable readings from the WMP. Why? The NK relies on the WMP and they work as a pair using interleaved data communications. So if the WMP isn't working then adding more complexity does not seem like a good idea.

[bernieL0max]

from memory last time I glanced at debug2 it seemed to be iterating through from -32,000 to +32,000 and then resetting... I might be mistakingit for another field though, I will head out to the shed and plug it all in again shortly

[bernieL0max]

yes, do what i see in debug2 is a cycle as described previously, counting from -32,000 to +32,000, then resettind and counting over again, I take it this is counting I2C errors?, and is probabl not a good thing(?)

[bernieL0max]

ok, so the pind to connect the Arduino aren't only around the eldge of the board... it looks like the I2C pins are inboard?... putting some pins in those now lol

[bernieL0max]

and hey-presto... things are working. I am seeing readings in MultiWiiConfig, I can arm it, the motors idle. if i bring up the throttle and start to move it about the motor speeds change and the yaw servo reacts. time to install the NK board !!!

getting excited now, but it just started bucketing down outside

[bernieL0max]

ok, so now with the NK installed I get an animated tricopter showing orientation, and the pitch and roll indicators respond well.

what I have now though is an LED blinking about once a second, and I cannot arm the motors.

defiantely a helluva lot closer than i was last night... any advice?

[bill516]

calibrate gyros, throttle back and left, pitch stick back (nose up)

[bernieL0max]

yeh, i was trying that with sticks... i might need to increase my endpoints on the TX... i did it by clicking the button in the config software and hey-presto, LED stopped flashing and motors armed with the correct stick inputs

[bernieL0max]

*proper* Maiden tonite... wish me luck

I say *proper* because last night in the dark i got a bit eager & crashed into the clothes-line... so we will discount that as being a pre-maiden test

[Olympian]

I recall seeing a post where the user test runs each motor while the FTDI is connected to the GUI.
He did this to compare the vibrations seen on the Z axis accelerometer to find unbalanced motor.
Sounded like a good test.
However, I am concerned about having the control board powered by the FTDI and ESC/BEC at the same time.
Don't want to damage my USB/FTDI ....
On same thought, having the ESC control cable plugged into the Control board means that 5 Volts from the FTDI is being applied to the output side of the 5 Volt regulator in the ESC at all times.... I have been doing this all along and the ESC's appear to tolerate this.
Does anyone have a definitive answer on how to power the motors while FTDI is connected to the GUI?

Olympian