Another Pixhawk fail here, "PX4IO not found",

[Marco Robustini]

I definitely don't want to discredit our board, but in a few months this is the third one that makes me the same problem,
identical.
Properly powered with the Power Module, all the connected ESC are OPTO and I have no servos on the output rail.
NuttX is active, I can get in, reflashing the bootloader but it says "not started" when I try to rewrite the firmware on PX4IO.
What can I check at the hardware level?
There's two 3.3V voltage regulator on the board? I see only one and I have already tested that it works.

https://www.youtube.com/watch?v=LQkoOEMgM0w&list=UUPcCCiqkWYZmvxVgHhDZDHA

Bests, Marco

[Craig Elder]

Marco

We first start by shipping you two boards were tested and were working when I sent them to you.

Of the two boards you returned to me, one had no issue and I was able to load firmware onto it and update the  IO firmware as well.

The second board had the 3v3 regulator for the IO processor blown.

I sent you back two working board with copter code loaded onto them.

There is no reason to re-flash the bootloader.  They were working when I loaded code onto the board before I sent them to you.

You told me on Skype that the 3.3V regulator was reporting 0.8V.

If the voltage regulator is dead the processor will not start.

I suspect there is some problem in your set up and you are damaging these boards.

It would be good for us to find out what.

Cheers,

Craig

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Marco Robustini]

Sorry Craig but i'm wrong, what seemed to me a voltage regulator is a shottky diode and it works perfectly.
If Pixhawk can fail so easily we must try to understand why.
I repeat, the board is powered correctly and there is no source of high consumption connected to it, only the receiver, telemetry and gps.
The problem is presented simply replacing the battery LiPo after yet another flight.
I've got five VR Brain installed on my drones and nothing ever happened to me at the hardware level, no single hardware issue, never.
I'm curious (it's my fault) and then to figure out where the problem lies, if I'm doing something wrong (possible but I do not think) or Pixhawk in certain situations becomes critical.
The curious thing is that the problem has always presented itself after you connect the LiPo, never in flight.
Now, I can't find in the pcb in the voltage regulator of IO controller to see if that's the problem, if anyone can help me, thank you.

- Marco

[Marco Robustini]

I have to follow these steps because here in Italy we certify our drones, and I want to try to figure out what hardware to use and its actual MTBF.

Marco

[Philip Rowse]

Hi Marco

    This is a strange fault, and not one that I have heard of before.  

    Was this straight out of the box? Is it a new pixhawk? 

    You say it has happened to 3 Pixhawks, were they purchased at the same time or close together? i.e., could they be from the same batch?

   do you have Eagle? you can look for the the regulator on eagle, if you open the schematic and PCB side by side, you can use the "eye" button to highlight the item on the PCB which should help in finding it.

Phil

Philip Rowse

Electronics Engineering Dept

3DRobotics

Ballarat

Australia

[Andrew Tridgell]

Hi Marco,

What receiver do you have connected, and to what port? Is there anything
else connected to the receiver?

One weak point I found a couple of months ago on the Pixhawk is that if
you connect a receiver to RCIN and the receiver draws too much current
it can damage the IO board. In my case I had a single servo connected
onto the receiver. That means the servo was powered from the receiver,
which was in turn powered off RCIN. That caused too much current to be
drawn from RCIN and it killed the board.

Cheers, Tridge

[Marco Robustini]

Thanks Tridge, this is very useful information.
So what's the maximum current applied to the RCIN?
Why the "+5V" in the rx rail (RCIN) is not direct connected to the Power Module to avoid these problems?
The receiver is "Futaba 7008 FASSTest", i've nothing else connected to the board that can consume a lot of current.
Which hardware may be faulty? The rx work fine.

Cheers, Marco

[Marco Robustini]

Hi Philip, thanks for reply, yes, i've Eagle installed now, I try to understand what happened.
Board are of different supplies, that faulty has the PCB green and is a revision 2.4.

Marco

[Philip Rowse]

If it's green, then it's possibly a prototype board. I would suggest sending it back to Craig again.

Before you do, please connect it up as it was, and take some pictures of your setup, exactly what is connected where....

Also, is it three faulty boards? Or is this the same board that you sent to Craig beforehand?

Philip Rowse 

Electrical Engineering Department

3DRobotics

Ballarat

Australia

[Craig Elder]

Phillip both of the boards that I just shipped were 100% tested and flown before I sent them to Marco.  Neither were prototypes.

[Philip Rowse]

 Hi Marco, I have chatted with Craig about this, and as it is a developer board which has you loaded a custom boot-loader.... I will let Craig deal with this.  

Philip Rowse

Electronics Engineering Dept

3DRobotics

Ballarat

Australia

[Craig Elder]

Just to be clear, one of the boards that Marco returned to me had no hardware issues.

Marco had loaded code onto the IO processor that would not communicate with the FMU.  Once I loaded the correct code onto it the board worked fine.

On the second board the 3v3 regulator supplying power to the IOP was blown.

In both cases the error reported was that the FMU could not communicate with the IO Processor.

Marco, let's take this troubleshooting out of d-d. 

[Marco Robustini]

Ok but it is quite strange, if I had loaded a wrong "IO Processor Code" the board would not work for a dozen flights with AC 3.2 RC9 and then he simply stopped working when changing the LiPo.
I try to do the task that Craig suggests.

Marco

[Marco Robustini]

The saga continues, we update: I went to the lab to do further analysis on Pixhawk that did not boot (red led), connect it in usb and... all is ok, start regularly!!!
I have no words, after hundreds of failed boot and without making any hardware or software operation now works fine!
Honestly this thing leaves me dumbfounded, how is it possible?
The electronics is mathematics, something either works or it does not work, can not work when they want, doesn't have a thinking brain... ;-)

Marco

[Craig Elder]

I have been meaning to respond to this...

>>>If it's green, then it's possibly a prototype board. I would suggest sending it back to Craig again.

the factory changed to green several months ago

--

[Philip Rowse]

Ah, ok, I will remember that... that's a pain, 

Philip Rowse

Electronics Engineering Dept

3DRobotics

Ballarat

Australia

[Marco Robustini]

As I wrote to Craig this board is booted regularly only when she was not fed for a while, just to be in thermal drift no longer boot, I've a video to prove it.
Now is flying to San Diego...

Marco

On Friday, September 19, 2014 10:26:45 PM UTC+2, Marco Robustini wrote:

[Craig Elder]

I have an interesting data point here.  A customer just brought in a Y6 with this issue.  It was locked up with the tones and the red LED that Marco reported.

I was able to update the firmware when it was powered off the USB but the Pixhawk would still give the same tones and red LED.  Pressing the safety button at startup did not succeed in updating the IO firmware correctly.

I decided to power the vehicle off the battery and this time there was enough current available to properly start the vehicle and the IO firmware updated correctly.

Everything is working and now even when the vehicle is powered only off the USB the tones and the LED behaviour are correct.  The whole vehicle is now working correctly.

--

[Paul Riseborough]

Craig,

I had one USB lead that didn't always successfully power up the vehicle as it would sometimes go into a mode with a red led (I threw it out). My theory was that some combination of inrush current on power-up and cable/sypply resistance was causing a problem. It might be worth checking the board voltage rise followoing connection. I don't have a storage CRO, so was unable to check this at the time I encountered the issue.

-Paul

[Holger Steinhaus]

Apropos hangup during boot: I have a self-powered hub here (DLink DUB-H7 Rev 2) that should be able to provide 2A. In some cases, the Pixhawk hangs up while playing the tones when powered via USB. This happens in may be 1 out of 20 tries. There is no other hardware connected, except one receiver (FrSky or Futaba FASST), a 3DR GPS module and a 3DR radio module.

I did not see a relation to a certain USB cable, it just happens randomly or not.

Holger

[Randy Mackay]

 

     It’s probably unrelated but there may be an issue with the I/O board firmware updating if the safety switch has been disabled.  It’s related to the I/O board not accepting firmware updates when it’s “armed” (i.e. safety switch has been pushed or disabled) because we didn’t want to allow updates inflight.

 

-Randy

[David Pawlak]

"This happens in may be 1 out of 20 tries"

And yet on others it never happens even with a standard conection.

That said. With my new computer I have never had an issue, and I'm sure it doesn't provide 2A of supply current. On a previous computer I had serious issues.

[Thorsten]

Yes, it must be the supply current. On one of my computer it depends on the length of the cable(s) I use. It sometimes refuses to boot and all I hear is periodic ticks from the buzzer. 

[Robert Lefebvre]

I wonder if this is the cause of my PX4 refusing to update. I am aware of the issue with power, so I did try to update while on battery power. But it still doesn't work.

[Jesus Alvarez]

Does it make any sense to disconnect the buzzer and see what happens? Maybe the external high power led too?

[Meier Lorenz]

Hi all,

USB delivers 500 mA, the buzzer draws much less than ten mA and the “high power” led is a high power led - it doesn’t mean it draws a lot of power compared to the USB power budget, also just a couple of dozen mA.

The #1 reason for USB issues are - surprise - flaky cables. It would be interesting to know if the results change if you use a good one (everybody cheaps out on USB cables, *except* the ones that come with Samsung or Amazon or Sony devices or other big brands - try one of that class). I did receive different ones from 3DR over time: The thicker cable gives 4.97V on the avionics rail, the thinner cable 4.87V. Now I measured this with a standard setup including radio, but *without* vehicle (which might draw more, if you have Gimbal, OSD, whatever connected), directly off a MacBook USB port (they know how to build decent hardware).

Now the USB spec allows a cable only to drop a maximum of 0.125V, so a drop of 0.13V is just outside the specs, and because Pixhawk doesn’t draw 500 mA during normal operation, it means the drop is much larger during boot up. A cable rated for 500 mA would only allowed to have 0.25 Ohms resistance, which is close to the 4.97V number I get with the better cable from 3DR (its thick and comes in a ring, not the tied up standard one).

In a cross-test my Kindle wouldn’t charge on the cable type dropping 0.13V - it kept switching on and off charging mode.

I did raise that point for a very long time with 3DR (to have the cables at least investigated, and replaced if my findings could be confirmed), and I hope this thread gives the incentive to source different cables to provide a better user experience. USB is complex enough that I don’t want to pinpoint every issue to this, but we really shouldn’t trip over bad cables, and my finding make a more thorough investigation into the cable resistance worthwhile.

Apparently other equipment manufacturers also got bitten by bad cables and put up these notes how to select a proper cable, which summarises the topic neatly:
http://www.gbs-elektronik.de/fileadmin/download/manuals/TN_Choosing_USB_Cable.pdf

Cheers,
Lorenz

P.S: The cabling on the front panel of your computer might also be flaky, and so are most USB hubs. You might try one of the backward ports directly off the main board instead. I have very consistent results with Mac hardware and/or decent USB hubs.

On 02 Oct 2014, at 16:00, Jesus Alvarez <wja...@gmail.com> wrote:

> Does it make any sense to disconnect the buzzer and see what happens? Maybe the external high power led too?
>

[Marco Robustini]

Hi Lorenz, in my case I don't think the usb voltage drop is the real problem, it does also powered by the the original 3DR Power Module:

https://www.youtube.com/watch?v=LQkoOEMgM0w&list=UUPcCCiqkWYZmvxVgHhDZDHA

Bests, Marco

[Mark Colwell]

Nothing like chasing your USB tail when things kinda stop. I power from a 7 port 2A external hub and trash poor cables. They cost more misery than savings.

[Mark Colwell]

Same for OTG to USB  cable for Galaxy S3, I found at GLB one with external power IN so I can connect it to charger when running DroidPlanner or other apps. No more weak power issues.

[Jonathan Challinger]

I have a special piece of plywood with a hole drilled in it and an xacto blade embedded so that i can rapidly strip all the delicious wire out of bad USB cables.

[Mark Colwell]

Nice DIY tool for a scrapper.

[john...@gmail.com]

>I have a special piece of plywood with a hole drilled in it and an xacto blade embedded so that i can rapidly strip all the delicious wire out of bad USB cables.

I have a special drawer at the office where I put the bad ones, for people who "borrow" cables and never return..