HELP!! AQ suddenly stopped working

[leopold1]

Ok, thx Jussi!

Looking for some solutions I found these ...
Opinions very welcome!

using only 1/3 of the pcb
https://www.sparkfun.com/products/8745

a 3,3V out ppm encoder
http://paparazzi.enac.fr/wiki/MeekPE_PPM_Encoder_Board

Or use a devider
5V in
\
/ 2K
\
--------- 3.3V out
\
/ 3K
\
GND



Any more suggestions apart from this
viewtopic.php?p=9691#p9691

[leopold1]

Jussi,

I did some measurements on the PPM encoder and it outputs a clean steady 3,3V. At the most with all thechannels max 3,427V. I even plugged the battery a couple of time to make certain it sparked. But it never went over 3,427V.
The PPM encoder was powered by a ESC.

Looking at the voltages I can conclude its stable so i could be something else that blew the MCU.

I also had a Xbee connected with an explorer board. Power came from the ftdi port but on the regulator board there was 5V to 3.3V convertor.

Apart from this, nothing else was connected.

[leopold1]

Jussi, what about xbee through the ftdi port. Could that been the problem?

[JussiH]

Yes, Xbee could possibly also be the cause - one team member has managed to blow out the serial connection on his boards. I am doing some research into this, and we will post some recommandations on how to protect inputs from spike damage.

I´d like to say that this is pretty rare, considering how many boards has been moved both in and outside the team. Its not a "general" error, but rather something that can happen if you are unlucky.

Did you scope the voltages on your PPM encoder or just measure with a voltmeter?

[leopold1]

Good to know about the xbee.
At the moment I measured with Vmeter but plan to scope it again befor I reconnect the AQ board.
When will the recommendations be noted in the wiki?

[Kisssys]

I've been using Xbee's for a long time with no problems, they are 3.3volt devices also. Of course they are a connection from the outside world that can bring in static charges when connected.

[leopold1]

Well,
I checked my xbee 900Mhz that was onboard the quad and behold ....... its dead
No more connection possible with xctu and it isn't bricked.
If this is the cause of the blown MCU I really need to get some protection on the signal lines.

Any suggestions how to overcome this. We still need to get the power from somewhere. And even a dedicated 3.3V powersupply could still give spikes.

Perhaps a startup procedure would help. First battery, then RX and last the xbee.

I've been using the xbee explorer board on my arducopter for over 2 years now without problems.

Come to think of it, the MCU didn't blow when connecting the battery but just after arming. Or it was a coinsidence.

To be honest I'm quite afraid to hook it all up again. Meanwhile I allready ordered a Spektrum TX module with AR7000. So hopefully the RX will allreday be ruled out.

[JussiH]

I got my Sherrer LRS working today with a Futaba FF9 radio. PPM is 5V signal level on the TX side, but on the RX the level is 3.3V and current is <1ma.

A PPM sum signal is very low current, my sherrer RX still outputs a valid signal with a 3.3K resistor on the signal line. But again, I need more data on PPM sum recievers. In my case I have <1ma current going through the signal line, so that fits well with the 3.3K resistor. But 660 ohm and 3.3V will give you a limit of 5ma. At 5V you will need a 1K resistor to limit to 5ma.

That is the only PPM device I have - so I am gonna need some data from other people with PPM RX´s and access to a scope. Measuring with a voltmeter will not show you a correct signal level, it will always be higher than what the voltmeter shows - on my Sherrer RX the PPM pin shows ~2.7V on the voltmeter, but the signal shows 3.32V on the scope:

I found some indications that newer Futaba RX´s are 3.3V signal level, but in other places I have seen indications that signal level is 5V for PPM sum streams. So it most likely differs between RX´s and converters of different make.

Regarding Xbee, they are indeed 3.3V devices, and I have been using Xbee 2.4 pro on AQ for a year now without any problems. But it could be a good precaution to put a signal diode and some 100R resistors on the signal lines. (An FDTI cable has 100R resistors on the signal lines to protect from overcurrents).

Simplest way to dump a signal level is to pass it through a divider or a small signal diode, which will give you a forward voltage drop of 0.7V - other options are the mentioned SFE Logic level converter - its basically just 2 dividers and two small signal Mosfets. I will post some schematics, but I am struggling a bit with the clock...But I think that down to 1.8V is a valid high on STM32, so if you want to play it very safe, some small signal diodes will give extra margin for voltage spikes.

I am gonna take a look at the 3DR radios next - but I guess that they most likely are 3v3 logic as well.

I will say that in your case it sounds a lot like you had an inductance on the Xbee signal lines and it killed both the Xbee and the MCU. Or perhaps the Xbee packed it in and took the MCU with it. There is basically no way to know exactly what caused it, but the fact that Xbee died too, is as close to a "smoking gun" that we can come for now.

The best precaution to avoid inductance in signal lines is to twist all power lines (GND and Power) around each other as they travel from one point to another and try to route signal lines away from main power lines.

I twist wires on all my copters and have never had problems with blown out MCU´s and I hook my Xbee directly to the board (but I always route the serial cable away from main power)

So provided you can rule out 5V on your PPM RX, I think the most likely culprit was "something" on the serial, either a malfunctioning Xbee or inductance in the PPM or Serial wires. Go over your wiring and route the signals away from main power.

[leopold1]

Jussi,

Thx for the heads up.
Could you please give some detailed info on the signal diodes. Specs, type, ...
Lookig forward for the schematics.

[leopold1]

Ok, the board works fine again but I'm strugling with the following.
I'm following Jussi's advice and putting some protection on the signal line of the PPM convertor.
This in the form of a 2.2K R and and 1N4148 signal diode.
The signal should give me a drop of 0.7V whilst the R should should limit the current at 1,5mA.
Before I installed this I tested on a led and it works.
But when I installed it on the signal line of the PPM conv the voltage at the end of the signal line went from 3.3V to 3.5V???
I checked everything but cannot figure out why?
Could someone shed a light please?