You are receiving this notice because our records indicate that you purchased an airspeed sensor from 3D Robotics. Our ongoing product testing has turned up a rarely occurring issue that can cause poor performance of the airspeed sensor.
There is no further action required if:
You are not using an airspeed sensor with APM 2.5
You are using an APM 1.0 or APM 2.0
When using the APM 2.5 with a telemetry radio and GPS, the current draw of these two devices may cause inaccurate airspeed sensor readings. An immediate fix for this is to disable the airspeed sensor in the Mission Planner.
For those who do need to use the airspeed sensor along with GPS and telemetry radio, and are affected by this issue, you can return your APM 2.5 to us so that we may replace a single component on your board and remedy this issue entirely.
We will do this at no cost to you. Simply reply to this message, and we will issue you a return merchandise authorization (RMA) with instructions on shipping your APM 2.5 back to us.
We apologize for any inconvenience this may have caused you and look forward to continuing to provide you with the best quality autonomous products at the most affordable prices.
The 3DR Team
if the problem is real is what it is and how to fix it yourself?
On the RFD900 site have a look at the pin diagram.
When you got the RFD900 you probably had a jumper between pin 2 and 3 of the connector. That is what powers it from the TTL serial link of the APM. Just remove the jumper, and put a BEC/UBEC servo cable from an alternative power supply on those first 3 pins. That will give ground and power to the RFD900 without stressing the APM.
I tend to use CasteCreations 10A UBECs, but you could also use the UBEC from your ESC if it has enough capacity.
Also note that radios are only about 20% efficient. So at 1W transmit power, the radio will be using 5W of electrical power. That is around 1A at 5V. That is more current than you should be pulling from the APM (especially as it varies so rapidly).
At 25dBm it's not nearly so bad, but might be worth doing anyway.
Thanks for that. I've got a spare one of these http://store.rfdesign.com.au/3a-regulator/.
It probably makes sense to use that anyway as you suggest, and it will save me having to send my APM back.
While I've got your attention I did a photomapping flight with a lawnmower pattern that had very narrow distances of about 25m between straights (less than half the turning diameter of my plane, and less than the waypoint radius). On one occasion when it reached the end of a straight it rocked/oscillated (in the roll axis) half a dozen times or so before doing the turn. It did that once, but on a handful of times it oscillated just once before doing the turn. It looked like it was having trouble deciding which way to turn.
Once or twice it actually turned the "wrong" way. I guess if the end of one straight, and the start of the next are closer together than the waypoint radius then it might hit them both at virtually the same time, then just navigate to the end of the next straight.
There was a slight wind, and I saw the plane was crabbing at times, so if the turn was to the right and it was crabbing to the left, it may have been a smaller yaw for it to actually turn the "wrong" way.
I was flying firmware v2.73. I've got telemetry logs, and a forward looking video that shows the wobbles nicely if you're interested.
Just to be clear, if I leave JP1 open and I power both the input and output from my BEC, then this should not be a problem? Or are the IO pins and telemetry radio still powered through this diode?
Alternatively, I can cut the telemetry radio cable, and hook it up directly to the BEC power. I guess I can do the same with the airspeed sensor power cable too?
Or do I have it all wrong?
I power from my xBee900, ioboard, minimosd, ublox from the telemetry port. Never had a problem but should I be worried about powering it all from the telemetry?
It's an APM2.5.4 purchases 1 month ago. Also should I update the PPM encoder or should it already have the latest version?
Thanks for the replies. That sounds about what I had in mind. I think it's better to power the airspeed sensor from the APM as it probably uses the internal voltage to do it's airspeed calculations and this assumes the airspeed is also running off of that voltage. If we power the airspeed from a different source, these assumptions by the APM may not be valid and the airspeed calculation may be off. Or, I may be talking out of my arse :-)
But powering the power hungry devices from a separate source is probably where we should focus and should be the only place we need focus.
I will also not be returning my APM as I can't wait weeks for my plane to be ready to fly again. Shipping from and to South Africa can literally take months!
Can I check that separate power source means more than just a BEC powering the output rail?
I've got the APM power module and have a BEC powering the servo output rail (JP1 is open) as per instructions here http://plane.ardupilot.com/wiki/voltage_current-3/
Do I need two BEC's, one for the output rail, and one for the RFD900?
servoes are anyway powered by the motor ESC if you have the tiny jumper on the APM2.5 removed. the most power hungry device is 3DR and Radio control receiver. the best is that they are both powered by a seperate BEC. a good one can be bought for $10. a 5v/3A BEC would serve the purpose and can be powered by the balance port of the flying battery. GPS and airspeed sensors draw a meager 10mA so they do not effect Fuse F2.
@moglos, the servo bus current does not passes through fuse F2. when you connect ESC to servo output rail the ESC powers the servoes directly without going through the F2 fuse. but please double chk. that is my conclusion.
Thanks for your help guys.
It should be pretty easy, I power my servo rail with a BEC so I'll just connect a female to female plug from the servo rail to the RFD900.
If I do that do I need to cut the power supply wires in the regular telemetry cable?