Dual Buck Converters, need design help pls.

[Bartman]

Hi gang,

We were talking about supplying power to a 3DR Pixhawk unit and a traditional method is to use power from the flight batteries reduced to 5V via a regulator (aka BEC).

I'd be happy to do that but the common buck converter isn't exactly an example of aerospace manufacturing sophistication. They are surprisingly reliable considering they only cost a buck and are sold by the dozen on ebay but, IMHO, to be a top-shelf power option for a heavier heli's control system, there should be redundancy designed into the system.

3DR recognizes people will want to do this and user @jfro posted a link where someone diagrammed a simple dual-regulator arrangement

http://drones.webstanz.be/blog/pixhawk-flight-controller-power-recommandations

Advanced configuration

: triple redundant power sources (power module as primary , plus two backup BECs – instead of one- to power Pixhawk’s servo rail):

A simple Tie bus circuit can be used to make the secondary power source redundant ! (therefore the power module can fail, a secondary BEC can fail while the third BEC will take over). In this scheme, a simple MBR1545CT integrated circuit is used. This circuit takes two BEC on its inputs and outputs only of of the two BEC according to the highest voltage (i.e. if BEC1 outputs 5.25V and BEC2 outputs 5.45V, MBR1545CT will pass BEC2 and blocks BEC1).
Here a tie bus circuit wiring diagram and example realisation with the MBR1545CT integrated circuit and a 6 pin JST connector:

So the basic premise is to use two buck converters in parallel with diodes to keep one's output from affecting the other's but can we do better by maybe conditioning/smoothing the output with capacitors and using LED's to indicate the BEC's are working (since a failure of one wouldn't keep the flight controller from operating normally but would eliminate the safety net of having two in parallel)??

Anyone want to chime in on sizing and positioning capacitors or LED's in circuit?

Bart

 

[jfro]

For the PIxhawk, IMO, it's a 2 prong reason. One is redundancy of the power, but the other is to negate spikes. Here is a warning on the Pixhawk I took from one of their tech docs.

It is dangerous to power the Pixhawk only from the servo rail, especially with digital servos. Servos may cause voltage spikes (as shown on illustration below that shows the servo rail voltage on an oscilloscope when a single digital servo attached to a Pixhawk is moved rapidly ). The key thing is that the digital servo causes the voltage on the rail to rise above the critical 5.7V level. Above that level the Pixhawk power management will cut power to the FMU and the Pixhawk will reboot. If that happens when flying you will lose your aircraft.

With 4, 6, or 8 ESC's, buck converters, gimbal, & led lights and not using any servo's, I'm not sure a voltage spike is something to worry about. The part about Pixhawk rebooting is a scary thought, I just don't know what the reality is with my setup. I also wonder if one put retracts on the system, would this be an area that might cause a spike. Anyway, just added this cause I'm curious and in thinking of putting my Pixhawk on a new MR with 6s batts.

On another note, there are some 3rd party power modules, but most pre-builts I've seen are 90 amp. ECalc suggested max pull on KDE 3520's would be just short of 30amps, so I'm guessing 30 x 6 motors could top 180, which is highly unlikely, but still a fair bit over the 90amp. Anybody want to make a suggestion that has some Pixhawk experience?

 

[Bartman]

Jim,

APM users with dual batteries and high current draw have used one power module on one input leg of their power supply and then just doubled the displayed values manually. i actually have an APM 2.5 that I never used but I did make a high current power module for it and that was the plan, to put it in one battery loop while leaving the second battery without a module.

Bart

 

[Bartman]

@R_Lefebvre

have any opinion on this?

 

[jfro]

I will be ready later today to either pull my xAircraft SuperX or Pixhawk from one of my quads and put it on my y6 lw/ 6s batts I'm building today.
My first choice is my Pixhawk, but hesitant without figuring out the power.

Since I'm coming out of a buck converter and currently not using any servo's, I'm not going to worry about spikes. Next is just if I want redundancy to the buck converter. I've never had one fail, so not sure I should worry about it. I also have 5v out of my frame as I'm using the Trooper which has both 5v and 12 volt out of the frame circuitry.


Main Board pcb has...

8 ESC Distribution board
XT60 Connector
6050 mould for 32 BIT Brushless GIMBAL
LED 1W
LED module socket for ESC and Motor
Gopro AV singnal input
12V 3A BEC output & AV singnal input (1 way socket provides AVOUT /12V/GND for FPV, other three only provides 12V output function)
5V 3A BEC output & AV singnal input (1 way socket provides AVOUT /5V/GND， for FPV, other three only provides 5V output function)
Battery Voltage Tester


The diagram above using the MBR1545CT, is said to just take which ever has the higher current, and passes that on. ie. if one converter is 5.1v and the other is 5.3v, it will pass the 5.3v and if that fails, it will switch over to the 5.1v buck converter....

That being said, wouldn't this work similar to what we are doing when putting 2 lipos on our MR's



I'd assume if I dial both buck converters to 5.1 or 5.3, that the pixhawk would be receiving 5.3. (or conversely use 1 buck converter from my 6s battery and 5v from the circuitry on my frame). If one buck converter failed (assuming it doesn't short out the overall output) then wouldn't I still be getting 5.1 or 5.3 volts to the Pixhawk and keep it running. This seems to me better in that there is no switch or power interruption of the power coming to the pixhawk.

Sorry, I'm not an electrician, so thought I'd throw this out in the hopes someone might have an answer in before I have to choose the Pixhawk or SuperX as the FC for my y6 build using 6s batteries.