Servo slow function query Futaba 14sg

[Nick Stock 2]

If anyone has a definitive answer to this I’ll be surprised and pleased..

 

I have been testing my newly installed mechanical retracts in a new model.

 

I have chosen Savox servos, one for each wing/retract.

 

Savox SC-0252MG+ Standard Digital Servo With Soft Start 0.19s/[email protected] SC0252MGP 

 

On my Futaba 14sg, I have had to set the end points right the way up and down to get the throws I need for each retract to lock in their up and in their down positions.  
 

The issue is this - I used the transmitters servo slow feature to attempt to make the deployment action look more scale.  
 

Bearing in mind these are servos rated to 10.5KG, there was minor if not negligible binding with the nacelle cutouts (model is a mosquito). 
 

On the side that I had programmed the servo slow speed to be greater (for scale asymmetric appearance) the servo stalled and burned out - a very hairy and Smokey moment!

 

On testing the remaining working side, when I dialled back the slow effect, the motion was much much smoother.

 

My suspicion is that the servo slow feature is having an effect of reducing the torque somehow of the servo. 
 

Has anybody experienced anything similar or can a specialist opinion be offered as to any resultant hindrance to servo performance - and some theory as to why my servo burned out?

 

 

[Paul De Tourtoulon]

If they are proper 180° retract servos, you cannot slow them down with radio settings, only by reducing the voltage to the servo motors.

 

 ps, I only have experience with Hitec and Futaba retract servos,.

Edited May 10 by Paul De Tourtoulon
ps

[Nigel R]

My take, on mechanical retracts, I'd put a 'proper' retract servo (they have the correct range of motion to properly lock with no servo load) - and live with whatever speed they operate at.

 

You could go to air-up/air-down retracts, with a valve in the line if you want slow/asymmetric. That works pretty well. Fancy control valves are available that can do this in one single part.

 

I've no experience with electronic retracts. You can definitely get a sequencer to get one up before the other, not sure if that can do "different speed".

[Paul De Tourtoulon]

I have electronic ones, the all metal ones at 45€, on 2 different channels one 'normal' the other a 15% slow, that makes them go in and out at different speeds, I also have spring air Robarts, air up spring down, pinching the air tubes was a disaster, so I pull up or tour and retract them when you can't see them.

[Dale Bradly]

Possibilities:

Servo was faulty and would have gone anyway.

Servo didn't have enough torque for the load required. Depending on geometry, there is a lot of mass to lift by the servo, possibly more than you think.

 

Either way: Servo slow functions from tx have no realistic effect on torque.

[dave windymiller]

I used a servo slow using an sg14 on a Hotspot jet air brake.  The servo, an s9255 9kg digital futaba servo would not lift the airbrake when slowed but was fine at full speed (via a switch). This digital servo appears to have a overload protection as after landing, the partially opened airbrake did not open despite no airflow but shut ok when commanded and then opened ok when commanded.

 

I suspect marginal servo performance on a high load surface combined with lack of momentum from the servo motor/gearing when slowed?

 

I have since changed to a 25kg savox servo (without the slow function) and futaba 16sz. 

[Nick Stock 2]

Thanks chaps. Update here:  I can only speak for my setup but as someone who very rarely makes definitive comments - I am 100% of the opinion that servo slow was responsible for the burnout. This radio feature if really cranked up is basically massively restricting the servo motor from spooling up to its normal rpm  whist at the same time drawing more current in order to maintain its positioning.  This caused a stall in my case and poof.  I had the slow setting at ‘27’ on that channel which in practise had the effect of slowing the travel time from either end point to roughly 4 seconds.

 

I took the view the next thing to try was to put heavier duty servos in and bought a pair of MacGreggor 20KG HV digital servos.  I also felt that powering these servos would be safer to do from a dedicated battery and I had a 1600mah 2 cell life battery to hand that I hoped would be suitable. At 6V these servos are rated around 17KG - way more than before.  
 

I reduced the servo slow setting somewhat and after I carefully readjusted the end points, in order to apply sufficient pressure on the retract pushrods at either end so that the mechanical locks were robustly engaged, in these positions there was still a pretty hefty buzz.  
 

I put a multimeter in the circuit and measured the peak draw momentarily hit about 1.5A with the lowest draw being mid way through the servo travel at around 0.2A. The end points with the hefty buzzing held at around 0.8A even with me giving each retract a bloody good shake around to check the load which was unaffected - proving the system was working as intended.

 

I decided the final test needed to be a timed control test and did as follows:

 

Fully charge life battery.
Set a timer on my phone for 30mins.

Switch on model and start timer.

Every 3 minutes flicked the switch to cycle the landing gear in the other direction. 
Monitor servo temperature by touching occasionally. Same for battery.

 

I did this and there were no

issues whatsoever.

 

I then put the battery back on charge and the charger put 267mah back in to the pack to my delight.

 

I do have a 1000mah life but the weight is negligible considering I’ve already got a kilo in the nose.

 

Im satisfied now so in my books, job done. 
 

Lesson: careful with the servo slow function and test test test. 

 

[Nigel R]

11 hours ago, Nick Stock 2 said:

The end points with the hefty buzzing held at around 0.8A

 

Have to ask - do you mean each retract servo draws 0.8A with the wheels fully up and/or fully down? This seems quite a lot. What does the servo geometry look like with the new servos? Is it a full 180deg range of motion?

 

Servo slow was not directly responsible - these features only changes the rate at which the TX changes from up to down, and vice versa. What it will have done, is expose an inadequate servo - one which now had to spend far more time in the portion of the movement where the current draw was very high.

 

Some other servo characteristics (inherent to the servo's particular feedback amp) will apply when servos move slowly against a large load, and in your case may have helped tipped something marginal 'over the edge' so to speak (more obvious in a case like dave windymiller describes where the servo could no longer "get going" against its load).

[Nick Stock 2]

0.8A combined. End points through the radio are near the max as the radio allows in either direction.  Can’t say if it’s 180 degrees but probably not far off.
 

Interesting take and thanks - I still stand by my theory that the motors aren’t getting up to revs so current rockets in order to fight the then relatively higher physical load - beyond the servos capability given the added length of time they were being asked to perform at this level for.  
 

[RottenRow]

Surely mechanical retracts have positive locks for the raised and lowered positions? There should be little demand on the servos when in either locked position. 0.8A, even for two servos, sounds like a lot of ‘standby’ current to me.

 

As the current has been measured at about 0.2A at mid travel, the 0.8A measured in the locked positions suggests that the servos are being stalled to some degree at least. Reducing the end points slightly, to give a small amount of free play in the locks, should reduce this current with the units still being positively locked.

 

Brian.

[Nigel R]

Yes... but with a bit of "it depends". 😀

 

I'd say, given that electric motors attempt to be "close to" a constant speed device (and apply as much torque as they need to, to get to that speed), that you've simply had "too little servo for the job" - and caused the servo to apply all its torque without much movement. This is inefficient, and inefficiency = heat. So yes, it is the extended time with the very high current draw that did the damage, all through the excess heat that generated. As you say - beyond capability for added length of time.

 

If the servo could "get moving" at the ends of the movement (where the linkage / geometry is often more favourable), then it can "keep moving" through the difficult bit (see dave windymiller's airbrake). But this is absolutely indicative of a marginal servo. In dave's case, very marginal indeed, as it stopped and couldn't restart at the midpoint. In your case even more so as the servo gifted you some magic smoke.

 

If you can stand the graph, a DC motor behaves as:

 

 

Torque is basically linear to current.

 

Right hand side of graph - full (stall) torque, the current is at the maximum, the speed at zero, nothing is moving, but everything is soon burning and smoking, not a good place to be!

 

If the servo is beefy enough to get the load moving, then the speed can rise a bit, and the torque (and current) can drop, everything starts coming down to within the servos 'happy place'. 

 

Note the point of max efficiency - it is around 80% of "unloaded speed" with around 20% of maximum torque/current. Flight surfaces, being used all the time, we wouldn't want to be going past that max efficiency point (and probably sitting below it, as a safety margin, truth be told). Something like retracts, or brakes, with 'occasional operation', we can load up the servo a bit more (perhaps going closer to the 'max power', in the middle) - but never to the point where the servo cannot restart movement. That said, a retract or brake won't be spending much time in these heavy load situations as it is cycled just a handful of times in a single flight. 

 

Short version - use a big enough servo*.

 

( * way too big becomes inefficient - however, that's an easier problem to solve, and involves less smoke and busted airframes )

[Nigel R]

5 minutes ago, RottenRow said:

Surely mechanical retracts have positive locks for the raised and lowered positions? There should be little demand on the servos when in either locked position. 0.8A, even for two servos, sounds like a lot of ‘standby’ current to me.

 

As the current has been measured at about 0.2A at mid travel, the 0.8A measured in the locked positions suggests that the servos are being stalled to some degree at least. Reducing the end points slightly, to give a small amount of free play in the locks, should reduce this current with the units still being positively locked.

 

Brian.

 

It does sound a little high for a "locked" position.

[Nick Stock 2]

Well I’m glad my conundrum is getting such appreciated attention 🙂

 

OK - so once I’d got the new servos in and had checked the radio was able to let me achieve a solid mechanical lock at both ends with a fair scope for fine tuning, I did then experiment by reducing the endpoints in the radio to find the minimum travel whereby I was still achieving this and then added a bit more as a margin - but as you say Brian it surprised and disappointed me too slightly.  In fairness I may have been a bit hasty here and perhaps spending more time and finessing I could reduce the end points a tad more whilst achieving the same robustness. The buzz will still be similarly prominent as I found with the tinkering that I did do.  I think I’m probably pretty much though and trust the settings I have arrived at.  
 

I could have made a career out of the exercise if I wanted to experiment with longer arms on the servo and less extreme end points but my hunch in any case was that the pushrod has limited play in the directions other than that to operate the gear and was therefore likely to cause a new problem and drive me completely over the edge..
 

The configuration and robust operation of the mechanical units in this case are just requiring the servos to do a bit of extra work to take up any slack in the pushrods so that they stay firmly pinned to their extremities given the geometry I’m basically limited to. 
 

Either way, bothering to run the 30minute timed experiment to me is evidence that my setup is doing very well to combat the risks involved. 

 

Nigel thanks too - useful as a reference to have a fancy graph especially with some explanation!

 

The only thing not ideal here is that the servos are working hardest in their default positions (gear fully up and fully down) which is how they naturally will be spending the vast majority of the time, only getting a rest midway through the action.  In an ideal world this would be the converse.   

 

 

[RottenRow]

If you are happy with your current setup Nigel then that is fine. Using a separate battery as you are for the undercarriage should prevent any problem affecting the control of the rest of the model.

 

Longer servo arms wouldn’t help you, as the torque produced would be lessened due to the length of the arm, and the servo would be more easily stalled.

 

Digital servos drive the motor with a PWM waveform, not pure DC. This gives greater torque at low speeds, and even at zero speed (holding torque). This is the reason why digital servos buzz even at standstill.

 

Usually, at standstill, this peak current will be relatively infrequent and of short duration, but they’ll still buzz. Apply a bit of load and the buzz will change as the PWM drive signal to the servo motor increases in width to compensate.

 

The peak current to the motor will be many times the current measured on a meter connected in series with the input lead, which will show an averaged out current (even if that average is taken over a very short period of time). Also the electronics inside the servo will ‘smooth out’ the input current to a degree.

 

Your original servo could have had a fault from manufacture of course, but the slow stepping of speed caused by the use of an extended servo slow operation would have meant it was taking too much current for too long.

 

Brian.

[Nick Stock 2]

Just to add a flavour of context here - this particular ‘ARTF’ model has consumed a mere 100-150 hours of my time thus far.. At what point does love of one’s hobby become a serious diagnosis? 😬