JETI REX Receivers & UDI

Dear V-Bar Team,

I know it's probably not so much your problem but I thought it may be worth asking for your opinion on the following:

http://helifreak.com/showthread.php?p=6723026

In the HeliFreak Thread, one can see an example of a heli that was a full loss due to problems with Jeti REX receivers and UDI (in this case with Neo). The problem seems to be occuring with other flybarless systems as well (in fact, Spirit just released a FW update to fix it).

From what I know, UDI is a Mikado invention and therefore all JETI Users may rely on your judgement as you are the experts. Please have a quick look at the video and the behaviour of a REX receiver (new, fast processor with more integrated memory) vs. an "old" ordinary EX-Receiver.

I'm sure all your VBar/Jeti users will appreciate your comments.

Thank you & BR,

Markus

Interesting. There is a Gap required between the packets, since there is no other means for framing. VBar needs 2ms minimum. Unfortunately there is no Value fixed in the Spec, since at this time, the timing was quite stable.
If the Gap is too small, there is the risk to confusion between only slower outputs or a real new packet, thats the reason for the 2ms

The shown behaviour seems to be sort of not right, since the transmission probably is not that random. I cannot see a reason why packets shall arrive back to back, or random.

Jeti has got a VBar (back at 2012) to test everything with it, so they shall be able to verify whats wrong there, it shall be easy to fix for them.

I am still not convinced, that this is the reason for some mysterious loss of signals. I am pretty sure the problem in most cases is a bad connection or some added inteference into that connection.

If Jeti cannot solve this for you, we may contact them an find a solution here.

so long ... Uli
http://www.vstabi.de

Hallo Ulli,

Danke für die schnelle Antwort!
Dear Ulli, thanks for the fast reply

MarkusD

Just a small addition since i read some comments on several locations about this.

The statement, that the processor is "too fast" or another one is "too slow" ist total bullshit. Just arguing, no technical background! All Controllers of the last 10 Years are totally relaxed to handle this little data in realtime.

The gap is a required synchronization feature. Each packet has to be separated by a gap to allow to detect the start of a new packet. This is not any sort of being to slow, but a required thing to allow a instant resynchronization even if packages are damaged during transmission. Some UDI Devices do not sent all Bytes in a Packet back to back, so if the gap is simply reduced, this can lead to massive problems on other UDI units (truncated packets)

The VBars input stage is trying to lock on the framerate of the input, this is needed to get a smooth and quick reaction. If packets arrive randomly, this lock probably will not happen. Thats not a real problem, but may degrade the control behaviour a little bit.

so long ... Uli
http://www.vstabi.de

Dear Uli,

a user of the jetiforum.de has just posted this mail from JETI Model

"new firmware is under tests just now. We added small gaps between data packets and
we didn't observe any problem with Spirit FBL. Public release REX firmware should be
available next week."

Did you had any contact and do you think that this is a potential fix ?

I belive that all users will appreciate and value your comment !

BR

Markus



1 mal bearbeitet. Zuletzt am 26.11.2015 22:31 von MarkusD.

No, no contact. I dont have any clue if it may be a fix or not.

so long ... Uli
http://www.vstabi.de

Hi Uli,

Thank you for your fast reply - this is however a bit worrying ..... As ordinary Human, I can't belive that they don't talk to the creator of the protocol......

We will see if Jeti can deliver a proper fix-but I would feel much better if they would have consulted your expertiese.

Thank you once again!

MarkusD

Hi Uli,

The Saga continues as Jeti released a New FW for their Rex Receivers - can you make a Suggestion what Timings Jeti Users should aim for?

Info Posted on helifreak:


Unread Today, 09:30 PM #75
TeroS
Registered Users


Join Date: Mar 2013
Location: Finland
Default
Quote:
Originally Posted by prototype3a View Post
Maybe it's just me, but I still don't feel like the problem has been fixed AND tested to verify that it was fixed. Sure, they seem to have fixed the packet spacing issue but I feel like no one even knows if this was the issue that was causing the crashes or not.
I honestly do believe it is based on two things:

1. My own readings with both oscilloscope and bench-testing with VBar attached to REX and doing a lot of shaking, steering inputs etc. And I mean a lot. I think I stirred sticks while moving gyro around for like an hour or so, always looking at the scope and VBar log, no sync errors.

2. This information from Jeti I got today:
Quote:
...We built UDI protocol support on SimpleSerialProtocol document from Mikado-Heli dated 1.10.2010...

...There is frame format description, values meaning and uart character specification, but no info about frame timing or gaps between frames in the Mikado doc.

...We proved that Mikado Mini V-Bar makes invalid servo output when there are too short gaps between UDI packets and Spirit FBL can suddenly stuck when doesn't manage more data.

...In REX fw 1.01 there are added small additional gaps between data packets and UDI period is also controlled via "Output Period" value. This solves problems with both mentioned FBL, but I can't say that it is primary REX bug.
I'm confident flying my 3000€ heli on REX & UDI now.

The sad fact is never the less that as we are moving from analog signals to more and more to digital systems with more code and processors between stick-input and actual servo-movement we are at the same time increasing the risk for signal errors, computing errors, protocol-mismatch and what not.

Wanna be safe? Don't fly. Wanna fly as safe as possible? Use some old protocol, connect the gyro to receiver with servo cabels instead of sum-signal.

But. There's always the gyro doing calculations.... Get the point?

Quelle:

http://www.helifreak.com/showthread.php?t=719210&page=4