HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Jul 20, 2012, 03:46 PM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
Discussion
Spektrum RSSI and improved data log

(Edit: Changed title to also include RSSI. Look at post #14 onwards for this feature)

I have found out that the Spektrum Data Log facility isn't of much use at all, as it is today:
Antenna fades don't tell much about radio link performance. Spektrum themselves recommend to only compare antenna fades relatively to each other. Yes, you can find out if one satellite position is substantially worse than the other(s), but you can't see if they are all bad.
Frame loss count don't say much either, especially on single-receiver equipment like the AR600 as it counts almost every antenna fade as a frame loss. And a hold (45 consecutive frame losses) is very bad, and should not happen at all.
So in reality, I can only see if my radio link is really bad (one or more holds). All the other values are of little to no use. After a flight I may see 100 frame losses, but were they spread out evenly across the entire flight, or have I had a few events where the frame loss almost accumulated to a hold?
I would very much like to see how close I am to the "really bad" situation, without actually going there.

Hence I am working on the solution: Put a small microcontroller on the data port of the receiver, and let it monitor consecutive frame losses. Then output a modified data stream with this information to be read from Spektrums data log display, or via a telemetry module.
I am not all the way there yet. I need to solder the hardware together. But I have written the microcontroller software, and simulations look encouraging.

What do you think? Would that be a useful tool, or do I just have too much spare time (yes, bad weather in the weekends will do that to you).
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Last edited by Mikael Pedersen; Jul 27, 2012 at 05:11 PM. Reason: Changed title to be more accurate
Reply With Quote
Sign up now
to remove ads between posts
Old Jul 20, 2012, 10:17 PM
Registered User
United States, AZ, Chandler
Joined May 2012
18 Posts
Mikael

I would question how practical this is (frame loss is not predictive except for very specific conditions).

You might look at the RF chip’s RSSI value- may be simple or very difficult, depending on your hardware, access to a logic probe, and your skills with SPI->MCU interface.

Many people have developed libraries you might leverage for SPI comms including RSSI- here is one reference:
https://sites.google.com/site/mrdunk...ystem-firmware
extrapilot is offline Find More Posts by extrapilot
Last edited by extrapilot; Jul 21, 2012 at 04:23 AM.
Reply With Quote
Old Jul 21, 2012, 03:21 AM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
I don't plan on reading the servo signals. I just hook up on the data line between the receiver and the flight log display or telemetry transmitter. Here I have access to the frame loss and hold information.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 21, 2012, 04:23 AM
Registered User
United States, AZ, Chandler
Joined May 2012
18 Posts
Sorry Mikael-

I misunderstood your hardware implementation- Ill kill that section of my post-

It might be useful to track several variants of loss (>20 frames consecutive, >10 frames consecutive, more than 50% loss over 100 frames, etc), and just output that in some sequential loop. Perhaps with that, you can narrow in on what is the better indicator for your style of flight vs link stability.

Regards
extrapilot is offline Find More Posts by extrapilot
Reply With Quote
Old Jul 21, 2012, 06:18 AM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
I am having the microcontroller keeping track of the 5 worst consecutive frame loss episodes, to be read out after the flight.
Also I will be monitoring the percentage of frame losses during a 45 frame period (1 second), and output that as a modified frame loss number. That in combination with a telemetry module will allow us to set an alarm in our transmitters if a certain threshold is exceeded.

One potential problem here. The original frame loss number is only designed to increment during a flight, where I modify it to display a kind of signal quality. I don't know how the transmitters will react when they see the frame loss number go down again. I hope they just dismiss the alarm and continue to work.
I have verified that the Spektrum Flight Log (SPM9540) don't have a problem with decreasing numbers.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 21, 2012, 08:24 AM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
Here is the preliminary schematic.
I will solder this together now, and report back later today.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 22, 2012, 02:54 AM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
After som debugging I finally got the software to work, at least on my desk. Now I am headed out to test it in the real world.

Here is a picture of my test setup, including in-system debugger and logic analyzer.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 22, 2012, 03:36 PM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
Success. It works.
I still have some more work to do on it, and testing it with a telemetry module. Perhaps also adding an extra feature, but when done I will make a webpage with the details, so you can make some, if you'd like.

I've taken a small video showing the data readout with a SPM9540. This is after a flight, and the receiver is an AR7610 with one satellite. The info A, L, F and H from the receiver is passed on unmodified, but B and R (which arent used by this receiver) is used to display my extended logs.

Datalog Enhancer test (0 min 46 sec)


First you see 27 frame losses, then 0 holds, then the voltage, then 255 fades on the internal antenna. Now we get to B. Here I display a kind of "signal quality" number. It is a percentage showing how many frames got through to the receiver in the worst second of the flight. Here it is 87%. I am thinking that this one-second time period perhaps should be a little longer, but as the criteria for a Hold event is one second, I just may leave it as it is.

The video then show L antenna fades, and then R. I use R to display a log of the 5 worst events during the flight.
Here it show that I've had an outage of 6 consecutive frames one time.
The next worst was 4 consecutive frames, also happened one time.
Then 3 frames, 1 time. 2 frames happened two times, and finally single frame loss happened 10 times.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 23, 2012, 12:53 PM
Registered User
rancho san antonio, CA
Joined Sep 2005
775 Posts
Very nicely done!

I was thinking of doing something the attiny 2313 and spektrum satellite, but wanted to
know exactly what these satellite send out (posted a question but received no answer).
I think I need a logic analyzer to snoop the data, so I'm thinking of getting a saleae logic
(the cheaper 8-bit version)

Are you using a saleae logic to do your snooping? If so, how do you like it?
mayday1 is offline Find More Posts by mayday1
Reply With Quote
Old Jul 23, 2012, 01:24 PM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
Thank you.

Yes, I do use a Saleae logic analyzer. It works perfectly. It is fast to hook up and the software really makes everything very easy.
I have an old (and huge) HP 96 channel / 500 MHz logic analyzer, with its own CRT screen and everything. But after I got the Saleae last year, I haven't touched the HP at all.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 23, 2012, 04:34 PM
Registered User
United States, AZ, Chandler
Joined May 2012
18 Posts
Mayday

The protocol was reverse-engineered several years ago, and there are not only samples available, but full source code for parsing. It is very simple serial data, easy to parse- easier still if you don’t have to write generic code to support 8+ channels and master/slave sats… Google is your friend.

http://www.rcgroups.com/forums/showthread.php?t=1284843
extrapilot is offline Find More Posts by extrapilot
Reply With Quote
Old Jul 25, 2012, 04:22 AM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
Progress report:

I have successfully used my unit with a telemetry module, and a DX18 as display. No problem. I still need to test the alarm threshold part though.

I am going to add an analog output, something along the lines of RSSI. That way I can hook it up to my Eagle Tree datalogger, and later the OSD when I get into FPV. I see this feature is highly requested, and that it has been attempted before with various degrees of success.
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 25, 2012, 10:21 PM
Registered User
rancho san antonio, CA
Joined Sep 2005
775 Posts
Quote:
Originally Posted by extrapilot View Post
Mayday

The protocol was reverse-engineered several years ago, and there are not only samples available, but full source code for parsing. It is very simple serial data, easy to parse- easier still if you donít have to write generic code to support 8+ channels and master/slave satsÖ Google is your friend.

http://www.rcgroups.com/forums/showthread.php?t=1284843

thanks extrapilot. are you trying to save me from buying a Saleae and use the
money for more RC parts instead :-)

I'm digging through the pages slowly.
mayday1 is offline Find More Posts by mayday1
Reply With Quote
Old Jul 27, 2012, 05:33 PM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
I have now added RSSI-like functionality to the hardware and software. It is not true RSSI, but rather a signal quality based on the amount of lost frames during the last either 1 or 3 second period (selectable).
The analog output is buffered, so it should work directly with the Eagle Tree eLogger. It seems to work fine on my desk, but I haven't tested it out in the open yet. And lacking an FPV equipped airplane, I may not get around to testing it for quite some time.
Apart from analog RSSI output, it can also display the signal quality as a number in the Frames Lost, just like above where it manipulates the B and R values. This can be sent with the Spektrum telemetry module, and with an alarm set in the transmitter, you may get an early warning when you are about to get into trouble. I don't know if this is useful, as when your transmitter has problems reaching the airplane, the telemetry module may very well have similar problems reaching the transmitter.

If anyone would like to put this module together, and test it, I'll gladly provide the software for the AVR. Especially if someone could test the RSSI part.

Here is the updated schematic:
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Reply With Quote
Old Jul 29, 2012, 02:03 PM
Certified Propellerhead
Copenhagen, Denmark
Joined Mar 2001
314 Posts
Doesn't seem like the fish are biting much here

I have tested the analog RSSI-like output with my Eagle Tree system, and it looks like it is working. Still haven't tested it with a real flight yet.
Here I am leaving the room with the transmitter in reduced power/range test mode. Then I do what I can to block the signal further, like sticking the transmitter in the fridge, or using myself. And no, I didn't sit on the transmitter.

The orange light is a satellite on the receiver, and the purple light near the center is a status led on my board. At the top center you see the RSSI value as Eagle Tree measures it from my board (named RSI).

Spektrum RSSI test (0 min 52 sec)
Mikael Pedersen is offline Find More Posts by Mikael Pedersen
Last edited by Mikael Pedersen; Jul 30, 2012 at 08:35 AM. Reason: Changed video
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
For Sale Spektrum AR7000 (New In Box) Latest Ver. w/Flight Log / data - $65 Shipped US whatasteele Aircraft - General - Radio Equipment (FS/W) 6 Dec 30, 2010 10:58 AM
Discussion Logging Spektrum 2.4 Antenna Reception Data atjurhs Sailplane Talk 11 Jun 14, 2010 01:42 PM
Sold Spektrum Data Logger Flight Log, like new never used viperjody Aircraft - General - Radio Equipment (FS/W) 2 Apr 12, 2010 12:21 AM
Sold Spektrum AR7000 New 75.00 Shipped Flight Log Data Port flynnaire Aircraft - General - Radio Equipment (FS/W) 3 Aug 05, 2009 03:22 PM
Wanted ar7000 or ar6200 and/or spektrum data log la528it Aircraft - General - Radio Equipment (FS/W) 2 Jan 09, 2009 05:34 PM