Mini servo comparison - my own test results
I thought I would share this in case anyone finds it valuable.
I recently bought a new glider. I have been advised by the supplier that it needs some very light servos for the rudder and elevator (also, I need to keep the total weight low). I’ve had several (cheap) servos break while in use on my foamies. Due to the expense of my new glider (Pulsar 4000e), I want to make sure I get good, swift, dependable and long-lasting servos in there, but not too costly.
To test, I bought:
- Spektrum A3020
- Futaba S3157
- Hitec HS-65HB (the karbonite one)
- MKS Ds450 (I thought I might use these for the flaps).
- Turnigy TGY-1800 (I had some laying around)
A quick play with the mechanics, un-powered (I'm sure we all like to force servos to feel what they are like ):
- Spektrum A3020 - felt really bitty, very bad, very noisy too; quite small amount of slop though.
- Futaba S3157 - feels cheap, possibly less slop than usual.
- Hitec HS-65HB - felt good: smooth and tight and little slop, if a bit stiff.
- MKS Ds450 - feels efficient, but not smooth and with some slop.
- Turnigy TGY-1800 - feels very efficient, but with slop.
All were powered from a Turnigy 5V, 5A SBEC (not great but will do for these tests), controlled by Futaba T8FG Super and R6208SB.
For all, re-centring after full displacement showed no observable offsets or faults.
The Futaba was the fastest from one end of displacement to the other. The others were comparable to each other - except the MKS which was shockingly slow (and got much slower when approaching the final position).
Slop while powered, with light mechanical loads:
- Spektrum A3020 - pretty good, but the motor is quite active.
- Futaba S3157 - not bad.
- Hitec HS-65HB - very good.
- MKS Ds450 – not so good, even though the motor is very active.
- Turnigy TGY-1800 - I didn't get this far as I did the harsh test (below) first.....
Very Heavy Loads:
I also did harsh testing, where I tried to place very heavy loads on them while powered. I was happy to test the servos to destruction; better to find out now than when on the glider....
- Spektrum A3020 - initially holds well; the cogs started slipping without too much load; got easier to slip when the test was reapplied. This is an indication of permanent damage!
- Futaba S3157 - held nicely (little additional slop), even with really heavy loads. However, the supply currents are high, possibly double the others with equal load. Also, the supplied servo arms flex noticeably, much worse than the others on test.
- Hitec HS-65HB - becomes surprisingly sloppy as load is increased. Heavier loads results with slippage, but this slip may be intentional, due to its 'Advanced Gear Train Technology'. I guess the gears are somehow spring loaded, which would explain the increased sloppiness with load. However, this does not get worse with multiple reapplications of heavy load.
I also noticed the internal switching frequency is around 50Hz. I thought I had bought the digital version; it turns out I didn't. Caveat emptor!
- MKS Ds450 - I could (comparatively) easily move the arm, not what I expect from the highest specced servo in terms of torque.
- Turnigy TGY-1800 - wow, that didn't last long. Started slipping, then within a few seconds slipped really easily, then started acting randomly - spins round and round, at random speeds, etc.
* For light loads, use the HiTec Karbonote.
* For all-round performance with heavier loads, use the Futaba, but be wary of the supply currents.
* Avoid the Chinese cheap crap
I will be ordering the slightly slower but stronger Futaba S3156, to be used with the CastleCreations 10A SBEC.
If anyone has another other comparative results to add (or other recommendations), then please do. Such direct comparisons could force the manufacturers to raise their game
. Servo - FAQ : Torque Measurement
with much more under sub sections:
"Servo - Alterations, Calculators, Databases, Leads, Repairs, Convert to an . ESC or winch & FAQ."
"Servo - Digital & Brushless Servo, Programmers and Software"
"Glitches & Jitter in Receiver, Servo & ESC - Causes and Cures"
"Radio Systems, Accessories, Alterations and FAQ" at
Alan's Hobby, Model & RC FAQ Web Links
My tests were to determine real-world behaviour and dependability. I was aware of the “Servo Database”. While it does correctly reflect the manufacturer's claims (some of which are obviously wrong), I think it doesn't reflect real-world quality, particularly slop and dependability (granted slop changes with use). Even though people do post reviews to it, they are usually not in direct comparison with other servos under the same conditions. This is why I bench tested.
For clarification: all forces were applied the same distance along the servo arm from centre; the results are comparative (if not absolute).
I take your point about the pulsed current and BEC overload. So to be sure, I repeated the tests using a powerful bench power supply and a rather large, local smoothing capacitor (0.1F should do it ).
The results were as before: the Futaba drew the most for a given stall load (peaking at 800mA average). The Futaba also held position the best, with the HiTec getting sloppy (and 'skipping’) and the MKS creeping (no point testing the Spektrum and Turnigy).
I have also done a dynamic torque test on the three. Again the current/torque relationships hold. Interestingly, the Futaba has a lot more stall torque than the others; yet it has considerably less dynamic torque; I guess there is some sort of mechanical lockup. The results show the Futaba cannot pull a dynamic torque of 0.9kg./cm – it is specced to be 1.5 but with what kind of torque?!? For once the MKS wins out here, sort of: it has plenty of dynamic torque but it never reaches the intended position. However, I have noticed stability issues – it can oscillate. It looks like its narrow deadband and significant slop does it no favours.
For the glider, I am thinking about using local, low ESR smoothing capacitors for each servo where installed in the glider. This should maximise holding torque (IR drops) while minimising system noise, at least when not at 100% duty cycle. This seems like a good idea to me (power up surge currents aside). I was thinking 2.2mF as it wouldn’t be a significant weight penalty. What do you think?
Your “FAQ Web Links” looks like it is useful. I will take a look.
I can already see that you have put a lot of useful info out there Alan. Any chance of a recommendation on a lightweight servo? Or one to test?
|Category||Thread||Thread Starter||Forum||Replies||Last Post|
|Discussion||Change Sun 10 blade 70mm (Test Results Consolidated)||901racer||Electric Ducted Fan Jet Talk||193||Feb 24, 2015 12:34 PM|
|Discussion||Anyone did the motor redundancy test for Naza Y6 mode? as my result, it failed...||Jason_cx||Multirotor Drone Talk||33||Jul 25, 2013 07:03 PM|
|Discussion||GoPro A/V out Delay Test with results||rcbif||FPV Talk||7||Jul 26, 2012 12:04 AM|
|Discussion||GPS Test Results||RC Man||Aerial Photography||33||Apr 08, 2012 05:45 PM|
|Krylon Fusion Over Solartex (my own results)||Derek Choice||Slope||23||Sep 18, 2003 05:06 PM|