Gimbals & Encoders 2nd half 2015 Have we Arrived?

[jfro]

Thought I'd start this thread to discuss the new gimbals and for people to share their coming experiences and video as these hit the markets.

The 2nd half of 2015 potentially looks to be what we have all been promised & waiting for the past few years with the brushless gimbals. Rock solid gimbals for gh4 sized cameras and down at a price that most of us can afford, ie, the $1,000 and down range.

There's also promises of new lighter cameras and at least one with 4 channels available for camera control from our RC controllers. Smaller/lighter/better quality is a given in the tech world, but RC control is a great bonus!

I'm going to start this off with my take on what I have experienced and read about so far.

Alexmos 32 bit controller. Works pretty good but I've never been able to get it to work day in and day out in a reliable and high quality mode. Have read others having the same problems, but some have working units.

Centerpiece. I have gotten great footage from this on a DYS Eagle eye with Gh2, Gh3, and Gh4. I have also had good luck on a DIY 5d style gimbal I made last year . I briefly tried it on another smaller/ lighter DIY gimbal without much success, but didn't have enough time to really have at it.

TPPacks has a model 300 with Centerpiece that appears to work with GH4 and smaller and the CP, but I have not direct knowledge other than what I've read on the forums.

All in all the CP seems to be progressing and when set up with a gimbal that works, it is capable of consistent and great stable footage. This I know from experience.

AlexMos 32 bit (AM) now has support for brushless gimbal motors with encoders. There is a cost advantage on the AM compared to Centerpiece (CP) even with the extra cost of the encoder motors, if these deliver as promised.

Another lightweight gimbal that is ready to ship soon using either the CP or AM controllers is here http://www.multirotorforums.com/thr...rmance-gimbal-for-gh4-a7s-size-cameras.24789/. This is called the Infinitymr. One interesting option the builder of this gimbal has is an option to purchase ($100 extra) the gimbal with encoder motors. I think the gimbal, encoders, and AM w/ 2 nd imu is $1095.

He has claimed that he autotunes it and flies. Says the holding power with encoders is rock solid, uses less power, and his video is very stable. Says, in the past (like many of us have experienced) with autotune, he'd get the shakes, now with encoders, he says it just works. If so, this is a big step forward. Proof will be soon as it coming to market soon.

Another promising gimbal I've been hearing about lately is the coming Gremsy H3. Rumored to be in the $1,000 range. I do not know the launch date on this, but let's assume the next 6 months or less. It appears to be another gimbal for the gh4 and down. Lightweight and affordable.

I believe we are just about there on the gimbal technology for $1100 and down gimbals. Smooth, jitter free footage is coming for our midsize cameras. All these gimbals are also lighter than what I've been using for the 5d/gh4 cameras. We have spent a couple years now chasing this technology for those who can't afford the Movi or other $$$ higher end gimbals.

Cameras are getting better and lighter too. The new Black Magic Micro camera with it's wonderful dynamic range, 10 oz weight (without battery/lens), Micro 4/3 format, and 4 channels of RC control for start/stop, f stop, zoom, and some other feature(s) that escape me right now, is coming this summer. If this camera delivers, even tough it's only 2k, it will be a great aerial camera with the potential to be flown on quads once we find a new light weight gimbal. I'm excited for both my X8 and my larger quad.

As a Hobbyist who strives for professional looking video, there are more flying opportunities with my quad. My x8 is a wonderful beast, but I get more flying and filming opportunities with a smaller quieter quad. DJI has proven that mid size quads properly designed can fly pretty stable in 15 mph winds or so with their Inspire.

We have heard the promises of easy to use gimbals and great footage before, but I actually think we might be arriving the 2nd half of this year for the $1,000 or so market. For the DIY gimal builders out there, I'd guess we will start seeing cost effective encoder based motors starting to it the market soon.

 

[MadMonkey]

I'm looking forward to reasonably-priced gimbals for personal work.

I don't know anything about encoder motors, what's the advantage?

 

[Motopreserve]

I've looked into the info on encoders on the motors - and there seems to be very little. It's still fairly new to the consumer market (although I think the high end gimbals have been using it for a while).

Great thread!

 

[dark_star]

will be interesting to see how these new AM boards and encoded motor combos work. encoders are the only way to get precision stability.

 

[Motopreserve]

Dark star, do you know how and why the encoders provide this accuracy? And how the controller divvies up the work to make it most efficient????

 

[jfro]

My understanding is the encoders give direct feed back to where the brushless motor is positioned. I think it's sort of like, when you power up and the gimbal levels itself, the motor then marks it's position. Say on the roll motor, the motor might say it's at 12 o'clock or say 0 degrees. This is probably most important for the yaw motor so after power up or adjustment, when it's straight ahead, it knows where the motor needs to be. If the IMU says the gimbal frame is tilted -6 degrees on the x axis, then it knows it need to go 6 degrees to get back to 0 degrees, or level.

Also, for some reason, I'm guessing the motor knows how to get back to 0 degrees with less effort, thus the power savings that Jakub alluded to in one of his threads.

Regardless, it seems the encoder technology may provide the next level that is needed for the gimbal market to make/help them keep our camera's level in a very smooth way. Also, it may help the AlexMos controller over the hump in regards to being a more reliable, easier to setup, and better gimbal controller. Which is all we really need at this point.

I don't think I'm out of line saying the the Movi has been the gold standard in Gimbals and that they use encoders on their motors. I think the demand for encoders will drive more manufactures to offer these and this always brings prices down. Win Win for us .

 

[jfro]

For me personally, the encoders will have to result in silky smooth video and simple autotune/setup procedures. I get great video with my CP. I'm looking for a smaller gimbal in the 1000g or less world and the decision to put an AlexMos or CP is really up to how the encoder based systems work if successful, the availability of the encoder motors, which by Jakubs shipping soon, I'm guessing encoder motors will be hitting the market very soon. CP is a great controller, problem is it doesn't work on many gimbals and it's expensive compared to the AM controller.

If this pans out, it will be a economic decision on the controller if we end up with a choice between two that work.

Over the next 90 days or so, we should start to see some results. As I said earlier, exciting times.

 

[Motopreserve]

No question the time are getting exciting.

Your description makes sense - and it certainly should help with power consumption of the motors do not have to rely on the PIDs alone to seek their position (shoot, then overshoot, then correct, then start all over again...).

I knew the Movi and Gremsy (and maybe DJI???) had them on their propriatary motors - and that they seemed to be working on those rigs. I found it interesting that Phobotic claimed it wouldn't help so much (if at all???), but it seems they are now working on integrating them into their firmware.

How is the connection made to the board? Are the current versions of the AM32 bit already capable of handling the connection?

 

[violetwolf]

Yes to my understanding the encoders plug into the rc inputs.

I am just talking to a well known manufacturer about building motors with built in encoders. I should hear back this week on their status.

Great thread! I agree with all the predictions above

UAV Media Pros

 

[jfro]

How is the connection made to the board? Are the current versions of the AM32 bit already capable of handling the connection?

$40 from Alexmos for encoder firmware.

http://www.basecamelectronics.com/encoders/

I don't know where I saw it, but I think you need two channels from your radio for the encoder motors. I'll see if I can dig up where I read that.

 

[Motopreserve]

I thought I had seen that too - but maybe you told me

If there is a fee - it's certainly firmware based. But the connection to the hardware could be dependent on the version of the board you own? Not if its the RC input - but I'm not visualizing how that works.

Just curious really. Not sure I'm into heading down the AM path - but I'll follow the progress with interest.

To be honest, I'm a little surprised that they charge for the firmware update/additional feature set. If this addition allows the AM boards to function consistently better, you'd think that alone would generate more sales of the controllers. The extra charge for the motors is going to scare some off. I find it odd they risk hindering potential sales with a small up-charge.

 

[jfro]

Anybody want to make a guess as to when Brushless gimbal motors with encoders will start to hit the market?

Anybody have any idea what the price increase might be for 50mm or 60mm motors?

 

[dark_star]

Without encoders there is no absolute position control. The gimbal control board and its ESC's only have motor pulse feedback which is very coarse, easy to corrupt and allows error. Encoders allow for a much tighter feedback loop. After all, that's what these gimbals are doing. It doesn't matter which one, everything from a cineflex to a tarot is trying to sense the smallest amount of error and correct for it as quickly as possible. The cineflex uses laser gyros and torque motors among other things and is essentially flawless in its angular error correction. A tarot uses cheap gyros intended for mobile phones and super basic brushless motors and it's angular correction is actually okay for the cost, but no where near the performance of the cineflex. The various other gimbals fall somewhere in-between. With encoders the feedback loop can be very tight (dependent on encoder resolution) and dramatically improve error correction. But it needs to be a fine tuned system. Just buying third party motors with an encoder output and hooking them up to a AM board is not likely to produce magic. The integration and tightness of the software/hardware combo is critical and not trivial to achieve. That's why I'm curious to see how these things end up performing.

 

[fmkit]

Anyone trying to zoom and hold position for more than 10s needs encoder to correct yaw gyro drift at least.
Any existing gimbal can be upgraded without swapping motors - just add compass
I've built number of DIY gimbals for high zoom camcorders, while zoomed in all the way there is no need to correct pan/tilt even if the gimbal left alone for hours

 

[jfro]

Without encoders there is no absolute position control. The gimbal control board and its ESC's only have motor pulse feedback which is very coarse, easy to corrupt and allows error. Encoders allow for a much tighter feedback loop. After all, that's what these gimbals are doing. It doesn't matter which one, everything from a cineflex to a tarot is trying to sense the smallest amount of error and correct for it as quickly as possible. The cineflex uses laser gyros and torque motors among other things and is essentially flawless in its angular error correction. A tarot uses cheap gyros intended for mobile phones and super basic brushless motors and it's angular correction is actually okay for the cost, but no where near the performance of the cineflex. The various other gimbals fall somewhere in-between. With encoders the feedback loop can be very tight (dependent on encoder resolution) and dramatically improve error correction. But it needs to be a fine tuned system. Just buying third party motors with an encoder output and hooking them up to a AM board is not likely to produce magic. The integration and tightness of the software/hardware combo is critical and not trivial to achieve. That's why I'm curious to see how these things end up performing.

I think many of us are interested in your comment

"Just buying third party motors with an encoder output and hooking them up to a AM board is not likely to produce magic. The integration and tightness of the software/hardware combo is critical and not trivial to achieve. That's why I'm curious to see how these things end up performing."

On the Infinitymr gimbal that is coming from Jakub, he published on one of the forums that he just put the camera on and did an autotune on the Alexmos software and flew. He says he is getting very stable footage. In it's simplest form, one might assume that it's that simple. this is very enticing as most of us with some experience know that we are not achieving that level of simple with great end results in our footage, (for me it's on the Alexmos, on the CP controller, yet I'm very satisfied on my heavier 5d gimbal).

Your comment about the integration and tightness of the software/hardware combo not being trivial to achieve would lead me to believe that Jakub's comment about "auto tune and go" that the AM integration into their firmware for encoders is handling all that and he has taken care of the hardware end with gimbal and motors.

It will be interesting to see what level of difficulty and what kind of success people have as the encoders get worked into our systems and if they are in fact, going to take us to another level of ease of use and quality of our aerial recordings.

 

[jakuban]

I think many of us are interested in your comment

"Just buying third party motors with an encoder output and hooking them up to a AM board is not likely to produce magic. The integration and tightness of the software/hardware combo is critical and not trivial to achieve. That's why I'm curious to see how these things end up performing."

On the Infinitymr gimbal that is coming from Jakub, he published on one of the forums that he just put the camera on and did an autotune on the Alexmos software and flew. He says he is getting very stable footage. In it's simplest form, one might assume that it's that simple. this is very enticing as most of us with some experience know that we are not achieving that level of simple with great end results in our footage, (for me it's on the Alexmos, on the CP controller, yet I'm very satisfied on my heavier 5d gimbal).

Your comment about the integration and tightness of the software/hardware combo not being trivial to achieve would lead me to believe that Jakub's comment about "auto tune and go" that the AM integration into their firmware for encoders is handling all that and he has taken care of the hardware end with gimbal and motors.

It will be interesting to see what level of difficulty and what kind of success people have as the encoders get worked into our systems and if they are in fact, going to take us to another level of ease of use and quality of our aerial recordings.

Yes. Integration between hardware, software and mechanics is crucial, it is the most important thing. Without rigid mechanics made from good materials you won't get good result. The same happens if you have the best controller but poor quality mechanics. I've put lot of attention to mechanics and rigidity of InfinityMR, and I think this is why I get such results.
When. It comes to Autotune. I run autotune for all 3axis and than separately for each axis to make final fine tuning. All videos I've posted ware done after that kind of autotuning. Of course I would say that such performance is possible on with encoders. Without them I've never achieved such high PID gains.

 

[jfro]

Jakuban, "When. It comes to Autotune. I run autotune for all 3axis and than separately for each axis to make final fine tuning. All videos I've posted ware done after that kind of autotuning. Of course I would say that such performance is possible on with encoders. Without them I've never achieved such high PID gains."

Do high PID gains not effect power usage? I think you said somewhere that you were using less total power to run your gimbal now. If I read that right, would that mean that maybe the power usage is more efficient and not so much being wasted with the encoders?

Do you have a shipping date set for your first batch?

 

[jakuban]

From the very beginning of using encoders I noticed both high PIDs and low power consumption. So somehow it happens that even with very high PID they don't affect power consumption in very visible way.

First units are going to be send on the beginning of next week and all units sold in preorder will be send before the end of May. There are some units available in first batch yet, if anyone's interested...

 

[jfro]

If Jakub's gimbal, the Gremsy H3 and some of the others coming out this summer are rock solid in their stabilization & ease of use/setup, what will DJI counter with. I would think that if the footage and use of all these gimbals were similar, and the new price was around $900- $1000, why would anyone pay the Zenmuse pricing for a gimbal that only fit one camera and 1 lens.

Might DJI come out with a $800 or less competitor? They have a history of being competitive when they want to dominate.

Gremsy H3 for flight is listed today at $895 @ 1.7 lbs... or 771g. This seems pretty light so am checking to see if this includes the vibration plate/setup.

 

[Motopreserve]

I think all the accessories are extra. That is definitely how the H6 is being sold. The dampener plate, quick release block, case etc are extra.