Current capacity of the 12V supply


#1

What is the Current capacity, off of the 12V-supply?

((((( the combo 12V/5V 4-pin male molex on the motherboard )))))

I am trying to drive a Velodyne Puck 16 from this, and sometimes, the voltage dips very low, possibly due to unexpected load from the VLP16


#2

There are a total of 4 supplies on the robot

12V Main
5V Main
12V Aux
5V Aux

The main supplies are shared with the electronics on the board. The aux supplies are solely for user use. Which is which can be seen on the pinout document here:

These supplies are all basically the same design. We say that these supplies can supply 7A of constant current. A couple of points to note:

  • The supplies are extremely robust. You can place a metal object in to the molex connectors and short out the supplies. The supplies will gracefully shutdown and then come back up again when the short is removed. There are similar protections for overheating etc. So I wouldn’t worry too much about your load damaging the supply.
  • The circuit is actually designed to provide up to 10A however running at this type of load for long periods of time causes the supplies to become too hot for our comfort which is why we backed it down to 7A on the data sheet.
  • They are switched mode supplies that buck the voltage from 24V through an inductor so this rather assumes that you are drawing a nice even current. If you try to suddenly suck out 7A for a hundred milliseconds and then go back to zero for a hundred milliseconds and then go 7A for the next hundred milliseconds the supplies will likely object due to the inductance of the output. We’ve observed this, for example, when placing a large packet radio transmitter on the supplies which intermittently transmits. If this is a problem it can sometimes be ameliorated by placing local capacitance on your power supply lines.
  • Our testing indicates that the issues with intermittent power consumption tend to be a problem if you have a highly intermittent consumption above 4A.
  • Important If you have a load which you think might cause the supplies to overload, consider using one of the AUX supplies - that is what they are there for - if an overload happens on the main supplies and that causes the supply to shut down it will affect the operation of the robot. On 5V main the raspberry pi and the logic circuits of the motor controller will shut down - on 12V main the motors will stop operating.

#3

This is a lot of great information, Dave.

UPDATE ~~ It now appears that the batteries had been discharged too much, to permit adequate current delivery to the 12V regulator. I directly measured the battery voltage off the terminals, and saw 21V !!! I’ve never seen it that low before. For the entire previous night, the battery charger was left “ON”. I can only guess that somehow, its connector was not properly seated in place. I am trying a second charge-cycle now.

QUESTION ~~ When external battery charger power supply is connected (through the 3-pin DIN), is the power regulator onboard the Magni flexible enough to simultaneously charge the battery, and supply the Raspberry-Pi / Motors / etc ??


#4

The system will work fine with the robot on while charging.

By design the regulated supply for the raspberry pi and the 5V circuits will keep operating to safely deliver 5V in pretty unusual power environments (how else will you be able to tell that you are in an unusual power environment).

For obvious reasons our safety and compliance certification assumes that the robot is not driving around while plugged in, but the electrical system is totally fine with it. This is why you might see in our compliance documentation somewhere - please unplug the robot while operating it - again this is a warning in the same class of “don’t operate in an oxygen tent”, “don’t operate in a bathtub full of water”, “this robot is not to be eaten” and “don’t pour 3 gallons of gasoline on the robot and throw a lit match on it”.


#5

Dave, I enjoyed your clever remarks, spoken against the headwinds of legal liabilities.

Anyways, after a little more investigation, it looks like our power connector . . . just was not plugged-in deeply enough to make adequate contact for charging the previous night.

I will admit the socket on the Magni seems mechanically “solid”. But, this may be a thing that could be improved in future.

  • Since the socket PC board makes direct contact with the motherboard, to avoid any stress to the main circuitry, I instinctively do not want to put too much insertion pressure on the connector, nor do I want to pull too hard to remove it. I believe that this concern interfered, resulting in an inadequate plug contact the previous night.
  • I would prefer to see this critical connector be mechanically separated form the motherboard, and provide the appearance of durability.

#6

The switchboard has been enhanced for the current production run but it is not yet a separate assembly rigidly attached to the chassis. Thank you for your inputs on decoupling physical chord yanking from potential damage to the main controller board. We are looking at things to make that assembly more ‘ruged’ and robust. We do what we can as time marches on to improve all aspects of the Magni in the order of the priority of each improvement or change.