DJI-WKM- The Gimbal Servo Saga- The final Results will surprise you.

araines2750

Hexa Crazy
This thread is about my "Saga" to find Gimbal Servos that worked well with the WKM.
About 2 months ago, I decided that I was going to add a 2 axis gimbal to my Hexa.
As such, I researched online for gimbals that would fit my rig as well as which servos others were have success & failures with.
I decided on a 2 axis carbon fiber gimbal which was to take (2) standard size servos.
At the time, I decided to use the servos recommended by the Gimbal Mfg. - (1) Savox SC-1251MG Digital & (1) Savox SC-1267MG Digital.
I got everything put together & attached to the Hexa. The gimbal worked...But...no matter wher I put the gains, the servos were either twitchy, skippy, pulsating or very slow and behond the aircraft movement.
I could not get them to work to an acceptable smooth & responsive manner. Some suggested that the servos needed more power and to add a second BEC. Not!
I then decided to try the gimbal mfg. OTHER recommended servos - (1) Futaba S9452 MG Digital & (1) Futaba S9551 MG Digital. They acted exactly like the Savox...No noticable difference!
At this point, I was unsure what to do, but I have an old (20 yrs old) Futaba S5101 analog servo sitting around from my 1st SR heli in the early 90's....Remenber the Kyosho Concept 30????
IT WORKED PERFECTLY!!!!!!!
No twitching, buzzing, skipping!!!!!
Since the rest of my OLD S5101 are still on my Xcell Gas...from the later 90's....which I still fly, I decided to order the current equivelant... Futaba S9101.
I installed the S9101 and.......It was skipping, but not buzzing or twitching....??????
The S9101 IS NOT the same as the S5101!!!
I then pulled a second S5101 from my Gasser to have (2) S5101 and....EVERYTHING WORKED PERFECTLY!!!!!!!
No twitching, buzzing, skipping!!!!! SUCCESS!!!!

$325.00 of Fancy, new digital servos.....FAILURE!
I had what worked best just sitting there collecting dust!
I have since gone on E-bay & bought (4) more S5101 ($28.00) so I can repoduce my results on other new machines.
I find it funny & frustrating at the same time.

DJI....You need to publish a list of servos that you have tested & verified that work well with your Fight Controllers- WKM & Naza!!!! And NO I do not mean the $$$ Zemuse...I cannot afford this!!!!

I wanted to share my costly findings with the Forum.
Amazing to me.. in this case Old (nearly antique) kicks the crap out of the $$$ new moden stuff.

I hope this helps others save $$$ and FRUSTRATION!!!!

View attachment 8893View attachment 8894
 

pwarea3

Member
araines
Great to hear you found a solution. I agree DJI needs to publish at least a few options that they know to work.

Do you have any video you can share?

Which gimbal do you have?

Thanks again for the info!
 

araines2750

Hexa Crazy
The Gimbal Kit I purchased was the last of it's kind. Made by HiGH+MiGHTY
http://www.dronesvision.com/high-mi...bal-for-gaui-dji-and-xaircraft-multi-copters/

They also make one for the Sony NEX5
http://www.dronesvision.com/high-mi...bal-for-gaui-dji-and-xaircraft-multi-copters/

I will make a video & post as soon as I can get the time.

There are all kinds of Gimbal "kits" available for Cheap to big $$. Direct drive, gear drive, belt drive...etc.
Personal choice will decide which you pick. The BIG issue is what servos you use!!! Do they work well with the WKM????
Mine is a small simple direct drive and now works great!
If you take out the $400 in worthless servos, It would have cost me just $178.00!!!!
I was able to sell the Savox to a buddy, Futaba are going on e-bay soon.
 

jes1111

Active Member
Glad you got the gimbal working nicely. The problem is nothing to do with DJI really - all that's coming from there is a PWM signal that's commanding the position of the servo - it may "wander" by a few µs but that's not what's causing the problem.

The culprit is the gimbal itself, which will have a considerable amount of mechanical "play" in it. Modern digital servos have high resolution ADC on the potentiometer and are aggressive in trying to hold their commanded position. The problems you've witnessed arise when the resolution of the servo is finer than the total play available in any given direction within the gimbal (including the play available in the servo itself). Put another way, you tell the servo to hold position "X µs" - the play in the gimbal axis and the servo itself adds up to, say, 10µs of equivalent movement - so the axis can actually move 5µs either side of the "correct" position. But a modern digital servo is able to detect that much movement as a deviation from its commanded position and therefore tries to correct it. Since that +/-5µs movement occurs, varying as the frame moves around in the sky, the result becomes a rapid "jitter" within the limits of the "play" - the system is effectively "bouncing" between the mechanical limits since they are greater than the measurement resolution. In the scientific/industrial world this phenomenon is called "limit cycling" and is very similar to the "oscillations" of a multirotor when it's over-tuned. As you've discovered, one solution is to use an analog servo whose measurement resolution and desire to correct even the smallest "error" is much less. The other solution, of course, is to remove (or at least reduce) the play from the gimbal/servo - through better design/construction of the axes, particularly the bearing supports. Since a servo is always going to have play in its gears (backlash) the only viable solution there is to "mask" the play with a belt drive between it and the load (a solution which can only ever be partially successful).
 
Last edited by a moderator:

araines2750

Hexa Crazy
Jes,
I respect your technical explanation, but the servos would jitter & skip when just sitting on the counter and hooked to the WKM- not even mounted in the gimbal or aircraft. Gimbal mechanics had nothing to do with the issues I was having.
Everything sitting stationary on a counter and servos still pulsating & jittering.
This is the only Video I have of how it "pulsed" / "twitched". Look closely, see the vertical "jiggle" about every 2 seconds.
The servo(s) did this even when plugged into the WKM & sitting on the bench.

[video]http://www.youtube.com/my_videos_edit?ns=1&video_id=aiwThO4Z2D8[/video]
 


jes1111

Active Member
The video demonstrates classic "jitter", whose cause can be attributed to the mechanical shortcomings, as I explained above. Yet you say the servo behaves this way when sitting alone on the bench being driven by the WKM. In which case, the only way to proceed would be to view the PWM from the WKM with an oscilloscope to see what's happening. Aside from the WKM just sending a junk signal (which is possible but I hadn't heard of it before) it could be that the servo lead is picking up interference and/or the signal level is too low and needs boosting. Either of these could be the cause and it's also credible that an analog servo would be less susceptible to such issues. With the digital servos, you could try twisting the wires over their entire length and also putting five turns of the wire through a ferrite ring, plus adding a PWM signal booster to the line.

In theory the response of a digital servo should be faster and more accurate, so it's up to you whether you want to pursue the best possible performance from that gimbal.

Good luck :)
 

DucktileMedia

Drone Enthusiast
I think you have way too many ear plugs on that gimbal! The idea is to use fewer of them so they actually absorb the weight of the camera. Just a thought.

As of yesterday, I am DJI FREE! It's a great feeling. I can actually talk to the guys who make my flight controller now! how about that!
 


DucktileMedia

Drone Enthusiast
I am absolutely sold on Hoverfly, yes. All around great company and product. The only way i will ever get another DJI product is if i am required to own a Zenmuse in the future but that would take a lot to go back to the dark side. But who knows, things change so damn fast anything could happen.
 

Kentjjl

Member
that is my video araines2750 has a link to my gopro was just mounted underneth to just see what it woul look like.there was no anti vibration fome used it was hald there with velcro.
 

questech

Member
I think you have way too many ear plugs on that gimbal! The idea is to use fewer of them so they actually absorb the weight of the camera. Just a thought.

As of yesterday, I am DJI FREE! It's a great feeling. I can actually talk to the guys who make my flight controller now! how about that!

I contacted DJI 8 times in the past 2 months for help and I have yet to receive a response :-(

My next High-end FC will Not be from DJI...

AM
 

Top