Pixhawk Power Supply Details

[Robert Lefebvre]

I'd just like to solidify the details of the Pixhawk power supply options.  I read through the blog posting and the ETH website, but have a few questions left. After these are answered, we should update the Arducopter website.

So as I understand it:

- the board is intended to be powered by the standard 3DR Power Module.

- due to the fact the system is actually 3.3V instead of 5V, the acceptable current range when using the standard PM is reduced to 60A.  Otherwise...?  It resets or burns out?

- there should be a new bespoke Pixhawk power module under development with a higher current and voltage input tollerance

- since the PM outputs 5V, and the system runs on 3.3V, there's obviously a 3.3V regulator onboard.  

What is the input voltage limit of this regulator?

- the board contains an ideal-diode controller, which controls the input power from one of two sources:  the PM port, or the servo power rail.  It has two inputs, but believe it will default to the PM if it can?

- therefore, and this is an important point, for multicopters, you can easily achieve power supply redundancy by putting 5V on the servo rail from one of the ESC's.  If the PM supply were to fail for any reason, it will switch over automatically to draw on the servo rail supply.

- the servo rail is rated at 10A, and 10V, so it is suitable for directly driving reasonably sized airplane and helicopter servos without need for a servo power breakout board.

- however, if the voltage on the servo rail is some unknown voltage higher than 5V, the redundancy is lost as the voltage is too high for the 3.3V regulator input.

What is that maximum voltage level?  

Does the ideal voltage controller also do over-voltage protection on both the PM port and the servo rail?  ie: if a regulator fails wide open as they sometimes do, the controller will protect the board from over-voltage?

[Andrew Tridgell]

Hi Robert,

> - due to the fact the system is actually 3.3V instead of 5V, the acceptable
> current range when using the standard PM is reduced to 60A. Otherwise...?
> It resets or burns out?

nope, it just reads 60A if you go over. The PX4-v1 could reset. The
Pixhawk has protection against over-voltage.

> - there should be a new bespoke Pixhawk power module under development with
> a higher current and voltage input tollerance

yes, I think so.

> - the board contains an ideal-diode controller, which controls the input
> power from one of two sources: the PM port, or the servo power rail. It
> has two inputs, but believe it will default to the PM if it can?

3 sources, the PM port, the USB and the servo rail. The PM is always
preferred if available. It will only use servo rail power for the FMU if
the voltage is below 5.7V. It can use servo rail power for the IO
processor up to 10V.

> - therefore, and this is an important point, for multicopters, you can
> easily achieve power supply redundancy by putting 5V on the servo rail from
> one of the ESC's. If the PM supply were to fail for any reason, it will
> switch over automatically to draw on the servo rail supply.

yes

> - the servo rail is rated at 10A, and 10V, so it is suitable for directly
> driving reasonably sized airplane and helicopter servos without need for a
> servo power breakout board.

yes, but see above not on FMU power limit for voltages above 5.7V on
servo rail.

> What is that maximum voltage level?

5.7V

> Does the ideal voltage controller also do over-voltage protection on both
> the PM port and the servo rail?

yes

> ie: if a regulator fails wide open as they
> sometimes do, the controller will protect the board from over-voltage?

yes, I believe so, but I don't know what the limit on that protection
is.

Cheers, Tridge

[Robert Lefebvre]

Ok, good info, thanks.  Is the processor/code aware of the status of the power supply?  ie: does it know failure states of the various inputs, which it's drawing from, etc?

So Gary, this should be added to the wiki please.

PX4 can reset if the analog voltage input goes over 3.3V (=BAD), but the Pixhawk is protected.

Maximum voltage to allow the servo rail to serve as a back up power source is 5.7V for Copters.

--
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/groups/opt_out.

[Andrew Tridgell]

> Ok, good info, thanks. Is the processor/code aware of the status of the
> power supply? ie: does it know failure states of the various inputs, which
> it's drawing from, etc?

yes. It knows which input it is getting power from, plus the voltage on
the servo rail.

At the moment we don't do anything with this information. We should be
logging it.

Cheers, Tridge

[Robert Lefebvre]

Ok, good, then we could make a new failsafe state where, if it has failed over to a secondary input, then could RTL.

Since I have your ear, do you know how to control the 4 output ports on the PX4?  There are two 500mA 5V outputs, and two hardware relays.  Is there a "pin" in the code that accesses these?

And I know the relays are gone from the Pixhawk, but are the outputs still there?

Cheers, Tridge

[Andrew Tridgell]

Hi Robert,

> Since I have your ear, do you know how to control the 4 output ports on the
> PX4? There are two 500mA 5V outputs, and two hardware relays. Is there a
> "pin" in the code that accesses these?

yes. They are documented here:

http://plane.ardupilot.com/wiki/arduplane-parameters/#Relay_Pin_RELAY_PIN

> And I know the relays are gone from the Pixhawk, but are the outputs still
> there?

I haven't tested the GPIO pins on the FMUv2 yet, sorry.

Cheers, Tridge

[Robert Lefebvre]

Oh, excellent!  So, should those pin numbers work in Copter 3.0.1, or will I need to modify our code to allow access to them?

This is actually looking really good for PX4, because I can use one of the relays as the trigger for my camera shutter LED system (Arduino Pro Mini, 5V trigger), and then still have the 3 PWM outputs for driving the AlexMos gimbal.  1 for pitch aiming, and 2 for pitch and roll stabilization signal!

Cheers, Tridge

[Andrew Tridgell]

> Oh, excellent! So, should those pin numbers work in Copter 3.0.1, or will
> I need to modify our code to allow access to them?

they should work for copter, as its a common library

Cheers, Tridge

[Andrew Tridgell]

Hi Robert,

One additional note on the PX4 pins. There is a note in the code that
the RELAY1 pin (pin 113 in APM) is currently used by the DSM/Spektrum
input code, so I think it won't work. The other 3 do work though.

Cheers, Tridge

[Robert Lefebvre]

Ok, thanks.

Oh, is all of this info relevant for 3.0.1 (ie: ability to use AC1 pin for trigger)?  Or is it recently updated (trunk)?

Cheers, Tridge

[Robert Lefebvre]

Hmmm... I don't see RELAY_PIN in Copter.  I hooked a wire up to ACC1, and it's pulsing constantly, about once every second or so.

[Robert Lefebvre]

Ok, so Relay Pin is only on 3.1 rc2 it appears.

I've been using the relay function for camera trigger on A9 with APM2, but I can't seem to get this working on Pixhawk.  I plugged into channel Aux 1, assuming that is the same as A9/RC9 in the code.  But, it doesn't work. Am I missing anything, or does this need to be developed yet?

Do you know if we are going to be able to get all Aux 6 channels working on Pixhawk?  Would be nice for the AlexMos gimbal control.

[Akshat4ever]

Hi rob,

The attopilot 90 amp C&V sense chip scales 0-89.4 to 0-3.3v...then there is no question of not being able to measure beyond 60 amps, unless the 3dr power module is using some other current measuring scheme... 

[Tom Coyle]

Hi Rob,

I am presently using three of the Aux Servo channels as digital outputs/inputs. I have channel 6 (virtual pin 55) as an input for the Start/Stop trigger function and channels 4 (virtual pin 53) and 5 (virtual pin 54) as triggers for my dual sonar configuration on my Pixhawk.

Regards,

Tom C ArduRover2 Developer

[Xerr Avon]

Hi,

This has been bothering me for a while, I have been using the auttopilot 180a because of this but then I have to use the servo rail for 5v. I also saw the wiki for using 6s batteries on pixhawk http://copter.ardupilot.com/wiki/common-using-a-pixhawk-with-6s-batteries/ they show a voltage divider on the output of the current sense. I tried this right away, but it did not work. after looking at the schematics of the autopilot and 3dr PM I see there is a 72.3k ohm (and a 110k on 3dr) resistor between the current sense line and ground which will change the resistance of one of the resistors in the divider. I think is why the wiki's voltage divider doesn't work, or didnt work for me anyway..

what I ended up doing is, using a 3dr PM, I replace the 110k ohm resistor with a 73.2k resistor and add another .0005 ohm shunt resistor in parallel to make it .00025 ohm this should scale it to 3.3v. It seems to work, at least in compassmot it looks like it is.

some one may want to look at the wiki, I am not 100% sure, but I am pretty sure the voltage divider solution will not work at least not as shown.

John

[Bernardo Diaz]

Hi Guys! I justa wanted to make a comment on the power module issue. I know that 3DR is working on a new one (Marco got one sample a few months ago) but MANY chinese cloners have new power modules which support up to 8S and up to 90A. So I expect 3DR to start shipping them soon

[Holger Steinhaus]

If run two chinese 8s modules here at 6s, with good success so far. Unfortunately these guys stick on their nasty molex connectors. They are quite easy to lose in-flight, better secure them with some hotmelt glue. Looking forward to some more solid stuff from 3DR.

Holger

[Xerr Avon]

 Hi,

The thing about the Chinese ones (at least the ones with 5v) I have seen is they are all scaled to 5v, for pixhawk and especially px4 we need 3.3v. I saw the 3dr HV PM's listed in the drop down in the latest ver MP, it looks like they will have a pixhawk ver and an apm ver so hopefully they will be available soon.

John

[Meier Lorenz]

This is not correct. The supply input for all PX4 family modules is 5V nominal.

-Lorenz

--
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<mailto:drones-discus...@googlegroups.com>.
For more options, visit https://groups.google.com/d/optout.

[Xerr Avon]

Hi, I am not talking about the supply, I am talking about the analog to digital inputs, the one most use that is labeled for current (actually it is on the px4io I think) is 3.3v max and I have read the board could reset if the voltage goes over. I have ben testing the PM I modified to be scaled to 3.3v.  what I did is put a shunt resistor (a.001 ohm) in series with the battery and measuring the voltage drop across it with a multimeter 1 mV = 1 amp. I tested up to 100amps and it is working perfect. even on the low end it is very accurate.

I decided to test it because I am getting higher current readings than I expected, on a hexa with Avroto M2814 motors, blueseries 40a esc's and 1238 props, strapped down, at full throttle I am drawing more than 180amps (compassmot flatlines at 180a now, with the unmodified 3dr PM it was flatlining at 60amps.  I stopped testing at 100amps just because it was kind of scarey getting close enough for me to see the meter :)

John

On Tuesday, July 22, 2014 4:19:19 PM UTC-5, Lorenz Meier wrote:

This is not correct. The supply input for all PX4 family modules is 5V nominal.

-Lorenz


Am 22.07.2014 um 22:52 schrieb Xerr Avon

 Hi,

The thing about the Chinese ones (at least the ones with 5v) I have seen is they are all scaled to 5v, for pixhawk and especially px4 we need 3.3v. I saw the 3dr HV PM's listed in the drop down in the latest ver MP, it looks like they will have a pixhawk ver and an apm ver so hopefully they will be available soon.

John

--
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<mailto:drones-discuss+unsubscribe@googlegroups.com>.

[Matthew Stamp]

Before I start inputting any info just curious as to who you guys are so I know who I am collaborating with.

Tyvm

[Matthew Stamp]

For a few ansers, just to reitterate, the pixhawk already has quite the internal power protection circuit. If any voltages or currents may or plan to exceed,even if there is a mechanism to help regulate that it would be wise to create a small external circuit to the input pin.

Furthermore, powering through servo rail is not the best way to connect the power supply so you have to really read into the documents and fish out a few notes/advice on cautions given. Being this is a large drone, finding eveyr little way to ensure maximum power effieciency is should be taken into consideration (will save heading, battery life, motor life and overall helpnot create internal circuitry forward currents through paths best to be avoided for our pruposes.

Any other power considerations can be handled by construction of small external power management circuit (assuming values arent extremely high the componets used should be nice and friendly to use on board.

Analog input resest for that minimum 3.3(V) would be bad. Would not be difficult to brach a BEC to deliver bare necessary power to remain uninteruptted or it would be possible to utilize a MOSFET push pull design to pull in voltage if detection of dipping for further compensation. If neither of these ides work there are power hat modules to summate problems furthermore.

On a non related note to your questions... keep in mind some processing chips may have max/min activation voltages/currents however ensuring proper operation again for max effiecieny (while completing the job) should try to be noted. (Ex. Raspberry pi I/O pins can casue excessive power dissipations and hence undesired heating if the value set is exceeded. 

Heres my sheets and documents this far. Update me which your and what you are currently trying to accomplish so I may stay in the loop

[Matthew Stamp]

I enjoy it much so. Very oriented toward the project I'm working on now

--
You received this message because you are subscribed to a topic in the Google Groups "drones-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/drones-discuss/yvHOWPx4nGU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to drones-discus...@googlegroups.com.

[jessie.m...@gmail.com]

Hi,

I know this thread is from a while ago, but I am hoping one of you more experienced Pixhawk users can help me figure out if I did something that would cause my new Pixhawk to fry.  My PX4IO board recently stopped working, and I want to figure out what exactly happened before I accidentally do the same thing to my next board.

For my tests, I had a servo in main out 3.  I was powering the Pixhawk with an ESC that was connected to an exernal DC power supply.  The ESC was taking in 11.1V and outputting a constant 5V to the servo rail.  I was also powering the board through the USB connected to my computer, since I was connecting to QGroundControl occasionally.  I had nothing in the designated power port.

Since my pixhawk was configured as a quad copter for these tests, I have since realized that the servo should have been connected to one of the Aux ports due to the update rate.  The servo did start acting twitchy while configured as quad copter, so I attempted to reconfigure as a plane in QGroundControl (I am programming for my custom plane anyway).  After that, however, my vehicle no longer connects to QGroundControl, and I cannot configure it.  I also cannot get into the NuttX Shell as I could before.  It appears that something has happened to my IO board (FMU board is still fine), but I am not sure what in my setup could cause that to happen.

My main questions are:
1) Is it absolutely necessary to connect a power supply to the power port? (sorry if this is a silly question, but I am new to this hardware and trying to figure out what went wrong)
2) Can having a servo on one of the main outs cause the IO Board to fry?
3) Does my setup seem bad?  Or could there be something else going on that is making the px4io board not be found.

Thanks, any help is greatly appreciated.
Jessica

[Akshat Deshpande]

This thread is dead and you should go to the new drones discuss forum after this. I have been using the pixhawk for a long time and if you were running the quad code and connected a servo to the main out 3, that would create problems for the servo because even when disarmed the pixhawk will be sending a 400hz 50% duty cycle pwm signal to the main 3 output pin, as this signal is meant for an esc , this high frequency signal will surely cause jitter on the servos, This will result in the servo stressing the regulator inside your esc supplying power to the pixhawk and the servo... it can sometimes cause failure of the regulator if the current transients caused by the jittering servo cause large losses in voltage regulation.... so yeah its a very wrong thing to power the pixhawk and servo with the same source... but even if your regulator shorted out and fed 11.1 volts to the pixhawk...even then it should not have caused the pixhawk to burn out...as the servo rail can tolerate voltages up to 20v, something does not seem right here, also if you power the pixhawk with usb then the pixhawk is not using the power from the servo rail as it prefers the usb 5v over anything else...so chances of frying your hawk are very minute ..     

--

[Randy Mackay]

I’m more software than hardware but I think it’s definitely better to power the pixhawk through the power port rather than just feeding the power in through the rear servo rail.  The problem that I’ve seen is that, especially if you have a servo on the rear rail, it can drag the voltage of the whole rail up and down a lot and I guess that could get through to the CPU (at least the low voltage drops would get through).  That could lead to brownouts but I don’t know how or if that could end up frying the board though.

 

-Randy

--
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.