Ive seen a few posts on the forum about power issues with the APM2.
I just had mine to the same thing when I was swapping input connections to a RX.
Ive looked into the schematics and the board and it looks as if there is a problem with the D1 diode that connects the power jumper / 5V Rail on the input connectors / board to the 5V rail on the output pins...
I think I have tracked down the specific device used and I have a concern... I believe this is the data sheet...
The problem is that the device is only rated at 30mA continuous forward current. Therefore if you are powering the unit via the output rail, the whole system cannot draw more than this (or a peak "surge" current of more than 500mA).
Has anyone else looked into or are aware this issue?
The other thing that Im now seeing is if I attach a standalone BEC to the 5V rail (via the input connections), the unit does not seem to boot up correctly. I have measured the output from the BEC and it seems to be 5.5V.
Does anyone know if this larger supply voltage (5.5V vs 5V) cause the unit not to operate correctly for some reason?
Can you explain the installation a little better (Zener diode to be installed in a reverse bias arrangement across the +5V rail)?
Zener diode does nothing for the problem. Zener diode is a shunt regulator meaning it shorts out the power source higher than 5 volts (over simplification but the basic effect). A shunt regulator works great on a snowmobile or motorcycle mageneto, but not exactly ideal when dealing with batteries.
In a real shunt regulator, a resistor is in series to limit current and the truth is, while it's great for low dropout, they also suck power like no other. Everything over 5 volts is burned as heat.
Zener should be viewed as a last resort.
The primary reason for the Zener diode is to protect the APM2 if the power to the board is reversed. It will clamp the voltage across the board and conduct enough current to cause the fuse to blow without being damaged itself. The secondary reason is that it will also help in an over voltage condition by clamping the +5V rail to a level that will not damage the rest of the board.
The primary reason for the Zener diode is to protect the APM2 if the power to the board is reversed. It will clamp the voltage across the board and conduct enough current to cause the fuse to blow without being damaged itself.
Umm, no, you would use a normal diode for that, you purposely reverse bias a zener for crude voltage regulation as I stated.
I get it, as a rather crude safety device. But back to basics, if the user cannot do surface mount repairs, doesn't the surface mount fuse that blows when the zener "protects" the board leave them in a the same "dead" situation? I mean if they couldn't replace the normal diode, they also cannot replace the fuse, or worse, bridge it and them make the same mistake.
The bottom line is that either the APM needs a switch mode regulator built in (AKA DC to DC converter), or else a bigger silkscreen of 5volts only on the board. You cannot protect against the stupid user without serious losses and further chances for failures.
Go back to the design and see the intent of the original diode that keeps failing was to protect the stupid user who cannot follow directions to not power USB and battery at the same time. This intentional protection caused failures. We blindly assume the newer complicated fix is less failure prone? I understand the desire, but sometimes, people need to suck it up and not change the design because a few people messed up. The diode failures happened because of user errors, not just random failures.
You cannot help these people, just let them make the mistake and maybe the 3rd or 4th APM they get the hint.
Let me restate this another way. So before, over volting the APM had a 50/50 chance of blowing the IC. Now, the protection garantess 100 percent the board will fail (fuse blown). If the user cannot replace the fuse or it cannot be bought locally and the APM doesn't come with a spare-the board is still dead to the user.
I'm having issues with my APM2 board that I don't understand.
• For a quad frame I am using a DJI flamewheel 450 kit with dji escs and motors.
• I have a external CCBec for powering the output rail since these particular DJI escs don't have builtin bec capabilities. I still removed the positive leads from each esc control wire however, leaving just the signal and ground leads.
• For a radio I am using a DX7 transmitter and a dsmx 7000 receiver
• I can power up the APM2 thru the USB connection and completely configure my unit with the Mission Planner successfully without issues. Everything functional.
vers GCS 1.19, firmware 2.6 mav10
Flight modes set to stabilize across the board
• When connecting the APM2 to a usb power source The POWER ON led turns on, the blue 3Dfix led flashes waiting for a fix, the gps lock led flashes. The armed led flashes waiting for me to arm the escs. This boot sequence takes a couple of seconds at most.
• If I connect a flight battery while powering the APM2 board via usb ( irrespective of a running GCS session connection)
I can arm and disarm and run the motors from my transmitter without issues.
• Now if I try to boot the APM2 from the output rail with my CCBec set to the recommended 5.3v.
My receiver powers on and binds on the input rail, I get the happy ESCs initalizing tones, the POWER ON led lights up, and the blue 3DFix led on the daughter board starts flashing but I don't get any ABC leds flashing ????
• If I measure the voltage on the input rail I see the proper .3 volt drop with the jumper in place. I am using the same flight battery as a power source for both the CCBec and Escs sharing a common ground through the battery.
Any ideas, why I can't seem to get the board to boot properly from the output rail? It just seems to sit there. I've gone through the wiki and there is nothing describing my problem. It would seem my diode is fine. I've tried reapplying the firmware and going through the config again, resetting the board via the reset button etc.
No luck. Any ideas?
Yupp, measured the voltage on the input rail with the output rail livened up. 5.3 output / 5.0 input.
Even compressed the jumper a bit to give a nice tight fit on the pins.
Like I said if I connect the usb the board boots and the ABC LEDs light up immediately.
If I use the output rail or input rail the board powers on with the green LED, 3dfix LED flashes as normal, my receiver binds, but the ABC LEDs don't light up. the RX/TX LEDs don't light up but I don't know if that is expected behaviour during the boot process when you don't have the usb serial port connected.
That would be me! :)
We will ship it back to 3DR for tests and repairs (or replacement)
Arducopter 2.5.4 had a known bug that it would not boot without reset from BEC.
Could you please explain what you mean by a "reset from BEC"?
From my professional experience, I am aware that some microcontroller will not "come out of a reset state" if the power supply rails are not within specified tolerances.
I am not sure if this is happening on the board or not. I have noticed that if the supply voltage is "high", say 5.5V, the board will not boot up.
If the supply is at 5V, the board operates correctly. I saw this when I attempted to power the board via an external BEC circuit (compared to the BEC on a ESC) and that the supply voltage was 5.5V not 5V.
Is this the type of Bug that you are referring to?
Arducopter 2.5.4 required that even when you supplied a correct 5 volts, the board often did not boot and required you to physically press the reset button. It seemed to work from USB power correctly without requiring a reset. This was a firmware issue not a voltage issue. It was fixed in 2.5.5 and also in 2.6.
What you are describing is why the instructions say 5 volts. Applying 5.5 volts was not recommended.