Frsky problem.

[mr.sneezy]

Just wish I had more time to dedicate to details like documenting and writing project pages.

I'm sure there will be a few here that will be happy to write some documentation for you when we see the design. You can just fill in the blanks.

[takilara]

Get on your *9.6 vdc line...select AC coupling on scope....50mv/div...no noise?

Tested now (still measuring directly on the battery terminals)
I confirm noise pulses, total amplitude approx 20-30 mV, 2-3 short pulses, then one that is about 3 times as long and repeat. (short pulses about 2 ms, long pulse close to 10 ms).
If i turn the Rx off, the noise increases.
No noticeable changes if i move the sticks.

However. Switched to Power Reduction mode, and noise more than halved. Noise amplitude in Power Reduction mode was approx 5mV.

I have one more possible test scenario that i haven't done. I have another identical Optic 6, that i could try the module in. but that Tx has no battery, so i dont think it would give us any more information unless i get a new battery anyways...

I'll check my LHS for batteries on monday.

[Ampair]

Tried it here with an Optic6, D8R rx (Telemetry), just unloaded servos on the bench
It seems remarkably consistent with individual servos. Hitec HS55, 3 tried smooth at all times; Hitek HS81 2 tried, jitter when moving; Hitek HS 311, 1 tried, smooth as could be wished; Hextronic HXT900, 4 tried, all jittered; Hextronic HXT 500, 2 tried, smooth; Vigor VS-12M, 2 tried, smooth as silk.
The Transmitter battery is around 4 years old, but still capable of consistently delivering 4 hrs plus flying time. It was freshly charged for the tests. High and low TX power did not seem to have any effect. 4 servos connected for tests. It may be imagination, but it seems that the jitter is worst on Ch 3.
Centring seems good except on the HXT 900s which have an intermittent buzz, apparently at random for brief periods of 1/4 second or less.
I didn't put a scope on the leads from utter laziness, maybe tonight.

[mr.sneezy]

Some other random thoughts.
I have some small paying jobs to get out of the way before I fiddle myself, but I think it may be interesting to measure the RX servo power noise, ripple on the 5V rail, then also measure the RX's internal 3.3V rail if possible.
Just a thought that maybe the issue is caused by peak power demand of the digital servos somehow, rather than what's going on at the TX end. Also I wonder if it's also worth putting a 1k resistor in the servo signal line to see if the drive from the RX is a bit weak for some servos, to see if it changes anything.
Martin

[Ampair]

Mr sneezy,
I just tried a "Y" lead - no effect. The jerky servos are still jerky whether there is one or two in the port, and the smoother ones are still smooth. That is an exact "jerky" btw, not a guessed one
Tried a different rx supply, still the same. Might try a low esr cap on the tx module.

[Ampair]

Now I have tried a stiff 10v supply for the tx.
Servos quiet; all of them. Might just throw a few capacitors around though (that will probably turn the power switch to charcoal).

[takilara]

I think i have a 10V PSU laying around as well. I'll test that.

What would be good size capacitor for a mod? (i'd hate to throw away a battery that still works perfectly with FM..)

[mr.sneezy]

Now I have tried a stiff 10v supply for the tx.
Servos quiet; all of them. Might just throw a few capacitors around though (that will probably turn the power switch to charcoal).

Hi Ampair. Inspired by the possible solution I just went and duplicated your test. The result was not duplicated though. No change. I also have 100mV of ripple on the 9.6V rail BTW (with only one TX pack).
I used an old breakout board I made years ago (forgot I had it) and used a second TX pack to just power the module separately to the TX. Unfortunately in my case the servo smoothness (or lack of) did not change. There may be two issues going on here, hence different results.
Cheers
Martin

Futaba breakout board with RF cable

MS254887.JPG (104.86 KiB) Viewed 15054 times

[Ampair]

Your mileage may vary - - this pack was on it's last legs about 10 months ago, so I stabbed through the jacket with needle probes and gave each cell a couple of jolts from a 470,000 uF capacitor with around 40 volts on it. Killed a few needle probes, and the pack danced about a couple of times. After that it got a dozen or so charge - discharge cycles at 20% C and seems to have regained it's old performance, but it possibly has a bit more impedance than originally. The FrSky module is taking a bit less current so I am still using it, which is probably stupid I know. I might just put 3 or 4 4uF7 tantalum capacitors around the transmitter on the grounds that if the problem(?) doesn't re-appear, I have fixed it. If the problem is back, it is obviously a bad batch of capacitors
It does not seem to affect models in flight - all I have tried it in is a Skyartec Skyfun, with no difficulties then a Multiplex Easy Glider that has two HS81s (down in the tail rather than long pushrods), and 2 HS 55s in the wing. It buzzed a bit while the tail servos were moving, but not really badly. In flight at a somewhat unwise distance and later at quite high speeds for an Easy Glider, there was no noticeable problems. I have just set it on the bench to try with the suspect transmitter battery, now at 9.8 v, and there is no jitter with any servo. I THINK I have everything set up the same as it was earlier except it is a little cooler now. (20.2 C, was around 22 earlier.) I may have less or more computers running than earlier; they surround the bench a bit. If this is going to be intermittent, I will be convinced that evil forces are at work again.

I was going to try switching to low power while the model was flying at a long range, but realised if it was out of range, I might re-set the failsafe to something novel. I usually consider these things while the glue is setting, but this time I stopped just before being stupid.

Tomorrow I may try to see if I can re-create the jitter, I don't know if I want to succeed in that. I will also try with a different FrSky module (non telemetry) but I have never noticed the problem with that module before/receiver. I'll also try with the telemetry transmitter and the older receiver.
Can you think of any other simple tests an old duffer can try?

[mr.sneezy]

Can you think of any other simple tests an old duffer can try?

Well not right off the top of my head, but I'll sleep on it and see what pop's in to my thoughts....
In the meantime here is a link to the cells I now use to rebuild my TX packs when they get unreliable. Thus far a very good cell, equal to Enerloops I think but at far less price. I just solder thin brass tags on them (very fast, with a large tip).
http://www.hobbycity.com/hobbycity/stor ... _to_use%29