Servos cogging

[jennifer6]

Hi Rob, I'm sure you will find where to put this question as it involves Eepe,Er9x,9x tx and Frsky module/receiver.
I have been flying with jr3810/frsky module and V8 one way gear now for 2 years without any problems. Recently (6mths) have changed to 9X tx and Er9x and have been flying various models without problem.
Yesterday I had need to change a servo in my flying wing, and ended up changing both(elevons) to HTX900 as I had them in stock. With these servos I now have a problem,-severe servo cogging effect.
Firstly I plugged my Frsky module into my JR tx and basically had the same problem but not as bad as 9X tx.Eepe has in the model setup software for er9x the ability to change ADC filtering, PPM frame Length and PPM delay.(default appears to be 22.5mSec, 300uSec and SING filter) And eepe suggests not to mess with these settings?
A message that originated from Frsky themselves suggest 18mSec and 250usec to stop servo cogging on some sevos.
I am using plain Er9x for one way, only are these settings different in the two way version of Er9x ?
Anyway changing these settings has not changed anything, have I missed something or am I confusing the issue.
Has anyone got experience with this problem and these settings, are they the default for a reason?, are there better settings, whats the downside of changing them.
I now have also checked my other models and there appears to also be a very slight cogging effect that does not concern me but can they be smoothed out.?
Any info on this would be really appreciated, Thanks JJ.

[wheelspinner20]

Can you explain what you mean by cogging? is this a regional term? Do you mean they flucuate rapidly or twitch??

Thanks.

Pat

[jennifer6]

servo moving in short steps, rather than smoothly.
JJ

[Rob Thomson]

I believe this is a known issue on the V8 series of modules.

FrSky apparently released a firmware update to fix this.

Check this thread on the GC forums.

http://www.giantcod.co.uk/forum/viewtop ... &sk=t&sd=a


Rob

[jennifer6]

Frsky cogging problem was with telemetry system only, apparently fixed with v2 upgrade.
Does not answer query on adjustments available within Er9x, to hopefully fix or improve similar problem I'm having with one way system, with certain servos.
Or indeed what are the correct adjustments of PPM signal and Filters.Or guides for these adjustments.

[jhsa]

I have frsky 2 way and never had a problem..

Just an idea. try asking here..

http://www.rcgroups.com/forums/showthre ... 8&page=402

[MikeB]

Is it just that the servo moves very quickly. The receiver output is a pulse every 20mS or so. If the servo can move from the last position to the next one in less than 20mS, then it will appear to jump. I believe the servo in question uses a coreless motor, this will probably make it start very quickly.

Does the servo cog when it is under a load? That would slow it down.

Mike.

[jennifer6]

Load does not make a difference. eg. If I hold the elevator by hand and operate the tx I can feel the movement presentlng jittery steps when moved slow or fast, held hard or soft.
I think this is a problem with these servos when used with Frsky, however I was hoping that the FW in Er9x will be able to overcome this.
The easy solution is not to use these servos.But now that I can see the problem, it is also slightly evident with all servos.
If I had not changed the servos in my flying wing I would still be a very happy 9X/Er9x user.

[MikeB]

I'll have a look into this by putting a logic analyser on the servo pulse output and seeing what it does when the stick is moved. It may be a day or two until I can set this up. Pat Mackenzie has an interesting, parallel port logic analyser he might use if he feels interested. Or I might download a copy and dig out my old laptop that has both a serial and a parallel port.

Mike.

[jennifer6]

Thanks Mike, Will be looking for any info.
I would love someone if they had the time to give me a detailed explanation of the pulse width, and timing functions in Er9x and what are the ideal adjustments .
And also the ADC filter functions and effects.
Unfortunately I am a total user and not a contributor, so its always a big ask to the people who really know to explain and inform.
Ideally this should be all in the manual, but here again needs someone to volunteer lots of time.
You should all be paid. Thanks.

[MikeB]

One test you could do is to set up a channel on er9x that is controlled by a switch, e.g.
CH05: +100% FULL Switch(GEA)
This should give an output that is either fully one way or fully the other, no stick A2D involved.
Now if you plug your servo into channel 5, does it still 'cog' when moving from one end to the other?
If it does, then the cogging is a function of the FrSky radio link.

Regarding the manual, if anyone has some paragraphs to contribute, just write them and post. I'm sure Rob will create an area for manual updates. It is fairly easy to add pre-written prargraphs to the manual. the OpenOffice source of the manual is on the er9x site.

Mike.

[jennifer6]

Thanks Mike, Tried that and although a bit fast, servo ran smoothly, great!
Am using V8JT one way module with V8R4 receiver.
Bound my DJT/V2 in one way to same V8R4 receiver and all results same.Bad cogging with Turnigy HTX900 servos in my flying wing.
Other servos seem OK, so I guess I keep away from htx900 servos(only used in very small models anyway).
Playing around with ER9x setting does seem to improve the cogging effect, but I,m worried about the "don't mess with these" in the manual.
Thanks again, regards, JJ.

[MikeB]

This makes me think this is definately a 'feature' of these servos. I think the problem is the whole radio system is geared around a PPM frame of around 20mS. So the servo only receives updated position information every 20mS. If it can move to the required position in less than 20mS, it will certainly appear to cog because it will move, stop, and wait for the new position information on the next pulse. With a coreless motor, these servos could well be fast, and fast to stop as well. There is an added problem of phasing between the er9x PPM pulse stream and the receiver servo pulse output. The receiver sends servo pulses at regular intervals, say 20mS, but er9x (and other transmitters) may have a frame length longer than this (er9x default 22.5mS). So sometimes the receiver doesn't get updated servo information between servo pulses and simply repeats the last one. Now the 22.5mS is so 8 channels of 2.06 mS pulses (extended limits on) leaves about 6mS for the 'sync' pulse. If you reduce the number of channels to 6, you should be able to safely reduce the frame length to 20.0 mS, or even as low as 18.5 mS (maybe even less if a shorter sync pulse is recognised, I think I found 4.5 mS minimum was needed). This frame length is less than the receiver servo update rate. It might improve the servo cogging, worth a test.

Mike.

[jennifer6]

Thanks again, Mike.
All very interesting. I'm going on a weeks fishing/camping trip up on the Tasmanian highlands, so will not be on "air" so to speak.
When I get back next monday will do some more tests and try and get my head around all this.
At this stage limits is turned off, the wing has elevons so its only using 3 channels, can I now safely reduce timing to 20 or less.
Actually I think I have missed something or got the wrong end of the stick, I'll wait till I get back.
Cheers, JJ.

[jennifer6]

Just like to put an ending to this thread.
After a lot of experimentation, with other brands of tx's , Frsky modules standard and telly, receivers etc.
I can conclude there is definitely no fault or oddities with 9X Radio, Er9x (any version) or Frsky module or hack installation.
The problem is just some servos(small percentage), and I must admit cheapies at that.Which caused the cogging effect I experienced.
I think Mike's reply got it right.
Thanks for everyones perseverance.
Cheers, JJ.

[MikeB]

We now have PXX working. This sends the servo positions to the Tx module as a number, not as PPM. When I sent the servo data to the module every 10mS, I reckon the servo response was smoother.

Mike.