Servo power-up movement

[David Germany]

Whilst investigating a series of crashes with a club members Wots-Wot (kit built) it seems the most likely cause was a momentary power interruption causing the servos to twitch on re-connection. Receiver and servo's have been replaced to no effect.

The cause of the interruption is most likely a faulty switch / harness or connection, and to this end we are replacing all wiring and connectors.

However this has got me to thinking why do some servos move temporarily on power-up, whilst some don't, and can anything be done to prevent this. It does not seem be brand specific, affecting Spektrum, Futaba, Hi-Tec and New-Power that I know of.

Obviously this is not normally an issue as we do not expect to switch the model off in flight but it would be interesting to know the explanation of why it happens.

[FlyinFlynn]

Are you sure it is the servos causing this startup movement? I would have thought it more likely to be a feature of the receiver not outputting the proper pulses while booting. Certainly occurs when connecting a servo to an arduino based system if you leave the arduino bootloader intact, it takes the arduino a second or two before it starts outputting the correct pulses.

[Cuban8]

David, are you attributing the crashes to the 'power up twitch' rather than the 'momentary power interruption'? I merely mention it as any sort of single twitch (glitch as we used to call it) will be very short lived and I'd have thought unlikely to crash a 'plane except under very specific circumstances. How long are you estimating the power loss to be? Did the model go 'free flight' for a few seconds and then flick into the deck? Battery power or BEC to the RX?

I wouldn't overthink why servos twitch at power up  - AFAIA, all my models' servos give a brief kick on boot up whilst the systems sort themselves out and gets into synch. Providing the servos are rock solid, centre and hold correctly with no spurious movement after the initial kick, all is well.  It's always been a feature going back to the old 27Mhz proportional gear that I started with in the 1970s - in fact the servos back then worked in much the same manner as current analogue servos do now - Very basically and in part, the servo compares an internally generated pulse with the command pulse from the transmitter (now transmitted fully digitally). It takes a fraction of a second for all this to synchronise and settle correctly at switch-on so the servos might give a bit of a kick before locking in.

 

 

 

 

 

Edited By Cuban8 on 10/10/2019 11:59:26

[David Germany]

It's the re-connection and subsequent servo twitching that ultimately caused the crash. On replicating the fault on the ground the ailerons would travel about 60% and the rudder about 20% in the opposite direction. You're right though about no servo movements when the power is cut, switching off the receiver and leaving it off the servo's do not move.

I was flying the model on the last occasion and it did feel as though something was wrong in the couple of seconds before the model flipped, my recollection was that the throttle was not responding, then the model rolled over and went into a dive, I then found I had control again, righted the model and cut the throttle hoping to glide back down.(it was only at about 15 / 20 feet at this point. Then the same thing happened but this time it just rolled into the ground.

[Cuban8]

OK, the model suffered a series of glitches as it headed towards the deck by the sound of it. If you can replicate the fault on the ground then the first thing to eliminate is any sort of problem with the power supply either battery or BEC. Check that the voltage to the RX is correct and not collapsing. Meter, LED battery tester etc.

How is the RX being powered?

Edited By Cuban8 on 10/10/2019 12:06:11

[Handyman]

I had a similar thing happen to me about 4 years ago when I purchased one of the new Black Spektrum Radio sets. It cost me two aircraft before I sent it away to be checked. Turned out that it was a fault on the RF Board causing the intermittent loss of control. My aircraft dived towards the ground on both occasions. I would suggest having the Tx and Rx checked out then go from there. My kit was brand new at the time.

[Cuban8]

I'm not completely ruling out a radio problem, but it's important to go through a number of initial tests and checks to eliminate essentially simple faults that could, incorrectly, be attributed to a radio issue.

[Denis Watkins]

Posted by Cuban8 on 10/10/2019 15:46:23:

I'm not completely ruling out a radio problem, but it's important to go through a number of initial tests and checks to eliminate essentially simple faults that could, incorrectly, be attributed to a radio issue.

Agree with Cuban

A good supply pack will hold the servos centred and still in relation to Tx sticks at neutral

As the voltage drops below satisfactory, the servos will operate wildly as described in the OP

I believe the Rx pack dropped voltage too low.

[David Germany]

Receiver was replaced after second crash, so unlikely to be the culprit. I'm thinking it's got to be along the battery / lead / connector / lead / switch / lead / Rx connection route. We checked the battery (2500 mAh 5 cell NIMH) after the crash with a load tester and it showed OK, and it still shows OK after 5 days so I don't think it's that but I will check for voltage drop with all servo's operating. I can't make the switch fail either even with lots of waggling.

As a matter of interest does anyone else get servo movements when switching on the Rx ? with the Tx on obviously.

[Denis Watkins]

Every time David, for years and years, servos move at startup

In fact, you would think there was a pilot in my Cub, testing the rudder and ailerons each time

[Cuban8]

Posted by David Germany on 10/10/2019 17:07:48:

Receiver was replaced after second crash, so unlikely to be the culprit. I'm thinking it's got to be along the battery / lead / connector / lead / switch / lead / Rx connection route. We checked the battery (2500 mAh 5 cell NIMH) after the crash with a load tester and it showed OK, and it still shows OK after 5 days so I don't think it's that but I will check for voltage drop with all servo's operating. I can't make the switch fail either even with lots of waggling.

As a matter of interest does anyone else get servo movements when switching on the Rx ? with the Tx on obviously.

Start from first principles, check that the battery is functioning correctly before doing anything else. To do this properly, please don't rely on a 'load tester' you've no idea as to what sort of load it might be applying or what voltage it's reading. Sometimes these devices don't put any load on the system at all. You'll need a multimeter, either a dial type or a digital one will do. Each type has its advantages. If you don't have one, it's worth spending a few quid to get one of each type - they're so useful. Cheap as chips on Ebay and although I doubt if they're as accurate as my own professional Fluke and AVO, they're perfectly fine for hobby use. Your 5 cell nimh should be able to maintain 5.5V when under full stretch from the models servos - all surfaces being moved rapidly for a minute or two. Note the battery voltage when the servos go crazy.

If you have a four button lipo charger, use its nimh discharge function at 2A and see how the battery's voltage holds up. You might have one or more dodgy cells. I mentioned that I found a bad cell in a pack recently in another posting - it does happen. Charge the battery normally and then do a 2A discharge cycle (set discharge cut off to 5V) Your fully charged 2500 mA battery should be good for around an hour before dropping to 5V. Much less than that and I'd suspect a bad cell(s). Hope this helps - if nothing else, to eliminate the battery as the problem.

All sounds a bit of a faff, but batteries can fool you easily, and as a result cost a few models before being identified as bad.

Edited By Cuban8 on 10/10/2019 18:00:38

[Chris Walby]

I am with C8 on this, had a few batteries fail and its normally under high load conditions when it becomes apparent.

Yes servos twitch on power up, some more than others but as long as there is full and free travel there should not be a problem.

PS I did have a UBEC and digital servos that would drop the voltage to the point the RX failed/rebooted, but it was found bench testing, The only time it did it was when all 3 servos were "stick bashed" at once, but if the model has been well flown then its a fault that has crept in.

[David Germany]

Fully charged the battery then discharged at 2A as C8 suggests, after 80 minutes battery was down to 5.7V so looks OK. Will try the extended stick stirring with meter attached tomorrow. I'm more & more inclined to think this was a connector / wiring intermittent fault so will replace the switch / harness and connectors and bind connectors with floss to secure. I think we are also going to replace the Rx anyway.

[Martin Harris - Moderator]

Probably way off the mark but you did select 5 cell on the load tester?

A friend asked me to fly his model a couple of weeks ago and I asked him to confirm that he'd charged the battery..."yes, charged it last night and tested at 99% on the checker"...

Two and a half circuits later it was a sad pile of splintered wood on the field. To cut a long story short, the battery was virtually flat and the automatic sensing battery tester had sensed the pack as a 4 cell...

Edit: Looks like you covered this during our crossed posts but well worth highlighting the possibility - could save someone a model in the future

Edited By Martin Harris on 10/10/2019 22:52:22

[David Germany]

Yes Martin, definitely selected 5 cell (6V) on the tester.

A good point though, would be an easy mistake to make.

[Cuban8]

Posted by David Germany on 10/10/2019 22:33:19:

Fully charged the battery then discharged at 2A as C8 suggests, after 80 minutes battery was down to 5.7V so looks OK. Will try the extended stick stirring with meter attached tomorrow. I'm more & more inclined to think this was a connector / wiring intermittent fault so will replace the switch / harness and connectors and bind connectors with floss to secure. I think we are also going to replace the Rx anyway.

Those results suggest that your battery is in excellent condition. Could conceivably still be a poor connection in the battery, but TBH that's very rare with decent branded packs. In your second post you say that you could replicate the bad servo movement on the ground - I'd monitor the supply voltage at the RX pins (unused connector or a Y lead) with a meter and see what readings you get when the problem occurs. If it remains rock steady when the servos are doing their own thing, then that suggests that switch and wiring to the RX is OK. Very risky to fly it as it is. Stick to a process of elimination for the time being.

Edited By Cuban8 on 11/10/2019 09:04:08

[Tim Ballinger]

Similar happened to me earlier this year. Flying straight and level , tried to throttle up, nothing. Then a significant twitch feeling like a flick. Followed by normal service. Landed all ok. Checked the Tm and saw significant Rx voltage drops (sbec) . Reasoned the ESC was faulty so replaced it. 2 flights ok then it happened again. This time not so lucky as I was close to the ground. Only bit I had not replaced was the arming plug & wiring. Perhaps significantly I could not find the plug at the scene of the accident (an external xt60). So I Again checked the TM . I could see no drop outs prior to the crash. Starting experimenting with momentary breaks in Rx power. Clearly a simple make/break was easily seen in the Tm, however if this was done sufficiently fast then the Tm showed nothing but you did get the servos doing the full movement associated with startup.

While not conclusive I decided the cause was with my arming plug and wiring having an intermittent contact. Possibly the missing xt60 indicating it was not in properly. Sounds unlikely, especially as I had been using it without incident for a couple of years.

However on the principle it is just another potential source of failure I have dispensed with arming plugs .
on reassembling everything (minus plug) all has been trouble free ever since.

So I highlight my experience not to denigrate arming plugs but to provide an example of how very short Rx voltage drops howsoever caused can both be difficult to find and , in the right circumstances, cause a crash.

Tim

[Trevor]

Whilst I agree that a power glitch of some sort is a likely cause, have you checked whether the failsafe settings are consistent with how the model behaved when it crashed?

[David Hall 9]

I've seen several instances of a poorly regulated/under spec'ed BEC causing problems. Under load, the supply can drop so much as to brown out the receiver. On resetting, several servos twitching can cause a further reset.

I had a cheap "no name" ESC on the bench with a cheap DSMX Rx. When it had three servos connected, any sudden movement of the sticks sent it into a reset. With digital servos, the momentary loads may be higher. I know of pilots that don't fit digital servos for this reason. In the air, these resets can look like a radio glitch, but a string of them (cycling resets) can give total loss of control. In my case the cure was easy, all the cheap Rx's and ESC went in the bin but I remain wary of using a BEC that is not up to the job.

[Nigel R]

"I know of pilots that don't fit digital servos for this reason."

Guilty as charged.