Replies to This Discussion

Permalink Reply by Adrian on June 11, 2012 at 12:26am

More Info:

After powering the unit up via the BEC (with 5.5V output), the unit does nothing for a quite a few seconds, then goes through the start-up LED sequence (rapid flashing alternatively red / blue then a slower alternatively flashing red/blue) then just gives me a single RED flashing LED which remains.

I grabbed another ESC with BEC (5V output) removed the signal wire from the connector and attached this to the input pins / rail.  The unit then did the start up sequence flowed by the flashing red  and blue as if it has started up correctly.

 

Is the APM2 that sensistive to the 5V supply rail? in firmware issue?

 

WIth either supply (ESC w/BEC or just the standalone BEC) the 3DFIX light was flashing which would seem to mean that the GPS unit is up and working and seeking sats.

 

Adrian

• ▶ Reply

Permalink Reply by Eric Tweet on June 11, 2012 at 12:37am

I've powered mine with a variety of different sources - UBEC, ESC BEC, etc, and haven't had noticeable trouble. Also, I've flown my Align TRex 600 with my APM board. This means that the board is powering 3 analog servos and a digital servo. Together these draw much more than 500mA peak (and I assume significantly high even continuous).

I have had power-up issues with my board, but it seems those bugs have worked themselves out. I haven't had power issues in quite a while. I hope someone can shine some light on this topic a bit, because the power capabilities of the APM have always been a question of mine.

Cheers!

• ▶ Reply

Permalink Reply by Adrian on June 11, 2012 at 12:50am

Have you always powered the servos / APM2 via the "output pins"?

 

Just looking at the schematic again, I have concerns that all the power for the APM2 / anything else connected (servos, 3DR Radios etc) are all powered via the D1 diode. If that fails (which it looks as if it has on a few peoples setup) that you will loose power to everything, which Im sure would not be pretty.

 

I have come up with a work around.. remove the Jumper and connect a lead betwen the 5V Rail on the output side to the 5V Rails on the IO Pins. This gives a direct connection between power rails and it would seem to be working ok as long as the ESC's BEC voltage is at 5V and not greater.

 

Adrian

• ▶ Reply

3D RoboticsPermalink Reply by Chris Anderson on June 11, 2012 at 7:23am

Adrian,

Yes, we are upgrading that diode, and new boards will be using a higher capacity one. It hasn't been an issue for most people, but there have been a significant number of cases where it burned out (about 50 of the first 1,000 boards) and had to be replaced, so your diagnosis is correct. 

Chris

• ▶ Reply

Permalink Reply by James Dinh on June 11, 2012 at 10:14am

Hi all,

I have had a similar problem with the output rail receiving power from an ESC.  It seems that the D1 diode in my unit has also burnt out and now I cannot power APM via output rail.  Word of caution, if you remove the jumper from JP1 and connect a link between the output rail to the RX (and via RX to input rail on APM) to power the board, you could cause problems to the RX.  I did this and my board powered up just fine until I noticed that my channels were starting to mix without changes to the TX.  When I throttled up on channel 3, yaw (channel 4) would value down, in other words yaw counterclockwise.  I found that the Futaba RX was receiving too much voltage (maybe a peak) from a linear 30 Amp ESC I was using to power the system.  I have now decided that the best route is to remove the jumper and power the system via stand alone switching BEC with 5V output.  I have not decided whether I should connect via RX or directly to input rail on APM.  My Castle BEC Pro actually has two independent servo leads going out, so would I be able to connect one lead to the input rail and the other lead to the output rail?  Can anyone lend some advice on this issue?  

Also, if all this doesn't work out, has anyone with similar problems had to send their board back to 3DR for repairs or replacement of the D1 diode?  How are they handling this issue, with quick or slow response, and is there a cost?  Please advise.

James.

• ▶ Reply

Permalink Reply by Dwgsparky on June 11, 2012 at 11:01am

Hi Chris

If this is a known problem then are we able to determine if a particular board will be effected. I dont want this to fail while in flight!.

Is then a prefered way to connect the APM , xbee, R/C rx ? At the moment my xbee takes power from the APM and my futaba Rx takes its power from the input rail, is this correct or am I waiting for the diode to burn out. My link is in place and I power  the APM from my 5v ESC to the output rail.

 

• ▶ Reply

DeveloperPermalink Reply by R_Lefebvre on June 11, 2012 at 11:22am

Hmmm... I was wondering about this.  I was actually thinking of jumpering the rails together just to get rid of the 0.3V drop from the output rail to the input, as my input rail only has about 4.7V which is a little bit low for my liking.

Now there's another reason.

• ▶ Reply

Permalink Reply by Dwgsparky on June 11, 2012 at 2:04pm

What if the two rails were jumpered together with a 2 core lead from the output to the input rail but with a diode in the positive lead. if the board diode burnt out then the input would still get its power without an interuption.?? Would this have any negative effects on the operation. We could easily fit a suitable rated diode in line. with a standard lead.

No difficult board soldering to be done.

Dwgsparky

 

• ▶ Reply

3D RoboticsPermalink Reply by Craig Elder on June 11, 2012 at 2:53pm

Hi Adrian

I have been investigating the power supply to the APM2 for the last couple of weeks and I appreciate you taking the time to look into this as well.

What you have actually identified is an error on the schematic rather than the PCB.

Diode D1 is listed as a RB751S40T1-D on the schematic, but it is actually a MBR0520LT1G.

The RB751S40T1-D has the same footprint as the MBR0520LT1G, but it has a 500mA continuous current rating and thus has an appropriate current rating for the powering the APM2 which is ~200mA. 

That is not to say the circuit is perfect.

We see three common failure modes with the APM2:

1. Unlike the APM1, the APM2 requires an external fuse when it is powered via the output rail.  The failure mode we see is that people are off by one pin when they connect something to the input rail and they short the +5V rail to the ground rail.    Because there is no fuse, diode D1 blows.
2. Overvoltage on the +5V input damages the serial multiplexer.  There is good evidence to suggest that transients (perhaps from the motors and the ESCs) are causing an overvoltage situation and damaging the TSA23157, U7.
3. Connecting the power supply to the APM2 backwards also damages the APM2, and generally it is the multiplexer that blows.

 

We have made some changes to the input circuit and will be incorporating those changes in the next revision of the board, however in the meantime I recommend you reference the 

Alternative ways to power your board 

section of

http://code.google.com/p/arducopter/wiki/APM2board

 

What are we doing about it before the next release?

 

1. As of last week, the MBR0520LT1G has been replaced with a MBR120VLSFT1 Schottky  diode which has a 1.0A  current rating.
2. Starting this week, APM2s are shipping with an external 500mA Pico fuse to install between the power distribution board and the APM2, and also a 6.0V Zener diode to be installed in a reverse bias arrangement across the +5V rail to prevent damage in the case of over voltage or reverse voltage.
3. The Pico fuse and the Zener diode will be made available from to anybody with a current APM2 who asks for them.

 

If you want to save yourself some time and get these components yourself, you can find them here:

 

http://www.digikey.com/product-detail/en/0251.500MRT1L/F3167TR-ND/7...

 

and here

 

http://www.digikey.com/product-detail/en/1N5233B-T/1N5233BDITR-ND/616

Cheers,

Craig

• ▶ Reply

Permalink Reply by Adrian on June 11, 2012 at 3:11pm

Thanks Chris for the info. I too have used a similar product with the same footprint when designing on Eagle.

Apart from this slight issue I am very very happy with APM2. I have it installed on my quad and I will be attempting the loiter feature today.. :o)

 

I have a couple of other questions.

 

When you mention the APM2 / multiplexer being damaged with overvoltage, what level of voltage? 12V? 5.5V?

As I initially observed, if the 5V rail is sitting at 5.5Vdc, the board does not seem to operate correctly (freezes?), but does not appear to be damaged.

 

Is there something that is that sensitive to the 5V rail being at 5V and prevents the firmware operating correctly?

 

I do like the idea of a crowbar circuit protection. Can you clarify what is the best email to request the mod that you are offering?

 

Thanks

 

Adrian

• ▶ Reply

3D RoboticsPermalink Reply by Craig Elder on June 11, 2012 at 3:13pm

reference this 

Alternative ways to power your board 

section of

http://code.google.com/p/arducopter/wiki/APM2board

• ▶ Reply

Permalink Reply by Sandesh Dhagle on June 19, 2012 at 9:27am

Hi,

I am also facing the same problem with powering of APM2. Will the following setup work with reference to 3DR quad frame:

Output rails powered from PDB's Power breakout for APM. As far as I understand the PDB use BEC output from Motor 1 & Input Rails powered from BEC of another ESC say Motor 2.

Or simply put a jumper cable between the input & output rails on +5V and Ground with Output rail being powered from the PDB.

I am not sure if this setup will have any adverse effect on the APM. Do I need to connect the ground of input & output rails?

• ▶ Reply

• ‹ Previous
• 1
• 2
• 3
• Next ›
• Page