Thread Tools
Apr 15, 2021, 04:28 AM
Registered User
After my experience with the X-Lite Pro and bad CNC digital gimbals, I happily bought the cheaper X20 with analog hall gimbals. At least if something goes wrong with them they will be cheaper to replace. My experience with them so far is that they feel quite good and I see no need to upgrade. They do have a bit of jitter visible on the raw ADC values, which makes me wonder how much filtering is being done on them (and how much added latency comes with that). It would be interesting to see the jitter on the digital PWM gimbals, but the reality is that the ADC stage has just been moved to the gimbal hardware and the only noise you are possibly eliminating is interference induced on the signal path from the hall sensor through to the microcontroller. Of course there is the possibility that the sensors themselves are just higher precision/quality on the digital gimbals, but again my X-Lite Pro experience with the MC12 gimbals was not stellar in that regard. So the jitter is probably the same unless the digital hall sensor is applying its own filtering before creating the PWM output. I would be surprised if there is any real functional difference either way.

Maybe people had better experiences with the full size digital gimbals than I did with the smaller X-Lite versions?

Ultimately we would need some tricky analysis to see what is going on with the sensor hardware and Ethos software and what kind of latency is resulting from it. I usually err on the side of “I’d rather be flying”, so I probably won’t bother. Also, the lazy approach here would be to guess that whatever is going on inside the Ethos black box is probably very similar to the OpenTX code as far as the stick inputs and filtering goes.
Sign up now
to remove ads between posts
Apr 15, 2021, 04:34 AM
Registered User
Correct

We all use our radios in so many different ways. For Ethos to entice Frsky users that have used Opentx for many years comfortably Ethos will need to at least attempt to match the possibility's in Opentx and allow for the amazing telemetry that Frsky/Opentx is known for and the diverse ways it is utilised...

I am glad of the work being done now, it will benefit me and my future x24.
Apr 15, 2021, 04:52 AM
Registered User
Thoemse's Avatar
Quote:
Originally Posted by locoworks
i think the difference in gimbals is plastic vs machined metal and not analogue vs digital as expected like the X10's
I've had many different TX by now. All sorts of gimbals. My personal conclusion: This is a a load of marketing cra*.
I would not pay for the *upgraded* gimbal unless I need the rotation feature. I don't care if they are digital or analog as long they work precise.
I just can't feel the difference.
Also keep in mind that gimbals in old high quality Tx were neither digital nor had hall sensors but they were great and still are.

I don't need BT audio wich is why I ordered the X20. The HD would've been interesting but there is not enough info avaialble and I am not willing to wait for it.

You milage may vary of course.
Apr 15, 2021, 04:58 AM
Registered User
Peter D Rieden's Avatar
It's a matter of semantics. Hall Effect sensing is an analog phenomenon, our transmitters send the information in a digital format. Somewhere down the data chain the analog data is digitised. Whether you put the A-D conversion inside or outside the gimbal subsystem in the system architecture is just an arbitrary decision at the functional partitioning stage of the design process.

PDR
Apr 15, 2021, 05:19 AM
Registered User
Quote:
Originally Posted by Oldgazer
Define accessible...
Damn, do I have to apologize to everyone individually for being my mistaken? Please refer to my previous reply to Mike. On checking my tx I realised that my statement was wrong for current versions. OK?

Quote:
Originally Posted by nheather
Indeed. And add to that

- what is the real benefit of CNC aluminium over plastic
- what is the real benefit of 10 bearings in the X20S over an undefined number in the X20
- realistically, what is the benefit of being able to adjust the throw, who will find that useful, who will use it
- realistically, what is the benefit of being able to rotate the gimbal by a few degrees, who will find that useful, who will use it

..................

I think in the real world, for average pilots, they won’t be able to tell the difference.

For me, I see the gimbals as being the primary difference between the X20 and X20S. In the UK there is a $120 difference between the two. So I want to know exactly what I would be getting for my extra $120.

Cheers,

Nigel
Bling and possible reduced flex in the yokes.
Bling for most of us.
I understand that there is a market from the multirotor race world(?) for reduced throws to help with rapid and repeated direction changes. Probably ticktock heli flyers too. I don't race and am not a hot heli flyer so don't ask me for more details
Mode 2 pilots with old worn out joints can rotate the stick assemblies to reduce control interaction between throttle/rudder and elevator/aileron. I'm all over this one!
As for bluetooth audio, I'm pretty sure other flyers find my current transmitter jabbering away to be a distraction at times so they will probably crowd fund me a bluetooth earpiece to suit.
Apr 15, 2021, 05:34 AM
Registered User
Peter D Rieden's Avatar
Quote:
Originally Posted by nheather
The CNC aluminium sounds obvious, there is the ‘duh, aluminium is better than plastic, because it just is, right’ argument, there is the ‘it is has more precise tolerances’ and ‘it is more durable’. But in reality there is no reason why plastic will be any less precise or less durable for the purpose it is being used for. In fact I have seen some users they they prefer the plastic gimbals because the aluminium ones feel more clunky.
Yes and no! Depending on the specific material, injection moulded plastics will all suffer a degree of shrinkage which must be allowed for in the tooling design. The amount of shrinkage can vary with material quality and process details, and of course injection moulded parts must have draft-angles on the side faces (or more, depending on how many impressions there are in the tool). This may not be important, but it can constrain the ability to have right-angled or parallel features in the art design which (in turn) can cause accelerated wear (this is a complex subject which I'm trying t summarise in a short paragraph, so I'm simplifying). Also injection-moulded plastics are ejected from the tool while still fairly hot, and in the ensuing cooling the associated shrinkage can warp the parts, especially if they have any long, thin curved elements. To some extent this problem can be addressed by using a thermo-setting plastic - something like Crastine, which is a powdered glass-filled polyester which cures by being pressure-injected into a hot mould. But that's expensive and I doubt it would be justifiable for these sorts of parts.

Against all the the machined part can be made with parallel/right-angled faces and has much greater thermal stability making tighter fits and tolerences more easily achievable but at higher cost due to the longer process times. At any given cost point the machined metal part probably has more limitations on part geometry than the moulded part.

In the actual finished part the moulded plastic item will have a higher coefficient of expansion and may also change shape with changes in humidity, so it has to be designed with wider fits than the equivalent metal part to avoid the risk of these variations causing binding or excess friction.

There are similar debates over the places where bearings are useful vs bling. Me? I have an X10s-Ex because I made a decision when I bought it. I have since handled a new X10-Ex and I can *just* notice the difference. If I had bought the X10 rather than the X10s I suspect O would have been hapy with it, but I definitely prefer the slick feel and lack of backlash in the metal sticks.

£0.02 supplied,

PDR
Apr 15, 2021, 05:45 AM
⚡⚡⚡⚡⚡⚡⚡
Quote:
Originally Posted by nheather
The CNC aluminium sounds obvious, there is the ‘duh, aluminium is better than plastic, because it just is, right’ argument

Nigel
https://www.rcgroups.com/forums/show...ostcount=10358
Apr 15, 2021, 05:56 AM
Registered User
Quote:
Originally Posted by Peter D Rieden
It's a matter of semantics. Hall Effect sensing is an analog phenomenon, our transmitters send the information in a digital format. Somewhere down the data chain the analog data is digitised. Whether you put the A-D conversion inside or outside the gimbal subsystem in the system architecture is just an arbitrary decision at the functional partitioning stage of the design process.
PDR
I agree with PDR here: analog vs digital Hall sensor gimbal is in fact of no consequence on the controlled process. For MC12P one of the feature is "Resolution: 12Bit overall resolution". To be honest is just marketing, you have this "12 bit overall resolution" also with pots.
What is true: Hall sensors transform variable magnetic field (due to relative movement between a magnet and the Hall sensor) to an electric voltage.
Pots do the same: they transform movement to a voltage. But pots may have "12 bit overall resolution" because the ADC in STM32 CPU is a 12-bit ADC and the pot voltage may have a "full range". But it's true you may have a jitter and also in time a wear on the pot. and it's also true that you may have an "ADC filter" (for instance you read 5 values , discard the extremes, do the average of the 3 or similar...just a simple filter, you may have very complex filter algorithms).
Hall sensor gimbals depend on the mechanical design, the size, the relative position between magnet and Hall sensor and so on. But you have to be aware that you will not get "full range" of values i.e, you will not get 4000 distinct readings (12 bits=4096 positions).
Just a few examples, some readings from OpenTX and some from EthOS:
Jumper T-lite (small Hall gimbals):
Ch1: 2AC to 68E => 994 values
Ch2: 0C1 to 4E0 => 1055 values
Ch3: 480 to 6A5 => 1152 values
Ch4: 1B5 to 44D= > 664 values
X-lite (a little bit bigger)
Ch1: 03A to 565 => 1323
Ch2:295 to 7C2 => 1325
Ch3: 2B2 to 7B3 => 1281
Ch4: 040 to 543 => 1283
X10Express in EthOS decimal values:
Ch1: 303 to 3610
Ch2: 650 to 3767
Ch3:: 645 to 3650
Ch4:315 to 3760
Jumper T16 in OpenTX:
Ch1: 15D to 465 => 776
Ch2:326 to 6E0 => 954
Ch3: 120 to 474 => 852
Ch4: 38E to720 => 914
Same Jumper T16 in EthOS (decimal):
Ch1: 690 to 2250
Ch2: 1650 to 3490
Ch3: 570 to 2280
Ch4: 1820 to 3650
What does a "digital Hall sensor gimbal"? It has an local MCU who reads the analog value from the Hall sensor, filters it and send it digital to the main MCU i.e. it does a part of the conversion that is already done in the main MCU. It does not "add" any resolution to the gimbal (you still have the same Hall sensor and the same magnet) and the reduction of work in the main MCU is minimal. You have to be aware that the main MCU does a final computation of values based on calibration values ..so in fact there is a minimal reduction of work (maybe the "ADC filter" part).
Do you have to care about this "reduced resolution" of Hall gimbals? Not at all. This 600-900-1200 range of values seems to not cover the 4096 "full range" but it doesn't matter at all. Why?
Consider that you have a full movement range of 4 cm on the stick. For 4000 values that means that you have a theoretical resolution of 1/100th of mm. Really? I have "sausage fingers", I'll be happy with 1/10th mm (400 samples).
You don't believe me? In OpenTX you have the "Analog" screen and for each stick and pot you have the absolute values (like above) and "percentage" values (between -100 and 100). Just a little exercise: try to do fine movements and move the sticks from -100 to 100, point by point. Try to get 200 distinctive positions on one stick, 40 mm divided by 200 => 5 positions per mm. QED.
As for the benefits of a "10-bearing gimbal" vs a "6 bearing gimbal" vs "4 bearing gimbal" I'll let our friend PDR explain it, I'm a cheap bastard and I see no point in paying 100 USD just for a gimbal gimmick (with 100-120 USD I prefer to buy a "full" RK TX18S "full options": Hall,MPM, touch screen) and my "sausage fingers" are quite happy with my current gimbals.
I do have Hall sensor gimbals on all my radios though ...
Last edited by ender67; Apr 15, 2021 at 07:12 AM.
Apr 15, 2021, 06:04 AM
⚡⚡⚡⚡⚡⚡⚡
Quote:
Originally Posted by ender67
I'm a cheap bastard and I see no point in paying 100 USD just for a gimbal gimmick
.....
Apr 15, 2021, 06:11 AM
Registered User
Don't get me started with this crap (or relative position of trims)...I can tell you that in my cheap T16 I have a couple of screws for controlling the angle of gimbals or something... I never bothered to use them. Gimmicks ...
Apr 15, 2021, 06:26 AM
Registered User
pumuklis's Avatar
Thanks to everyone for the help!
I replaced the original SD card in the X10 and reinstalled ETHOS.
IT WORKS!
I was also able to boot the R9 mini Rx.
I set up the mini quad for ETHOS and I have already flown with it in the apartment.
ETHOS still has some shortcomings on the old X10, but I hope they will be fixed.
Apr 15, 2021, 06:46 AM
Registered User
Thanks for all the gimbal information and opinions.

I certainly heard from the community when I was looking at the Jeti DS-12 that most couldn't not tell the difference between the metal and plastic gimbals and that some preferred the plastic over the metal because it felt smoother.

You guys are quite correct, Hall Effect is analog by nature so at some point there has to be some Analog to Digital Conversion (ADC), it is a question whether to do that on the Hall Effect chip or separately on the PCB.

As some has suggested it probably makes little difference. ADC is expensive to do at a very high quality, I can't imagine that FrSky are spending much on the ADC but likewise I doubt they are spending much on Hall sensors either. So any difference is more likely to be down to the interconnects. If done on the Hall chip, the analog signal does not leave the IC whereas if the ADC is done on the PCB then it must travel through flying leads, connectors and PCB tracks - as such I'd say that the analog Hall with PCB ADC is more prone to interference and therefore jitter.

So where I am coming from is that the 'enhanced' gimbals cost $100 ($120 in the UK with sales tax). I was assuming that the X20 > X20S was analogous to the X10 Express > X10S Express. In both cases the primary difference is the gimbals, the better gimbals on the X10S Express cost an extra £54 and the better gimbals on the X20S cost an extra £87. With the X10S Express the gimbals added bearings, CNC aluminium construction and digital Hall sensors. It is sensible to assume that the X20S gimbals would add the same.

But if it is that the X20S gimbals are still analog (not confirmed yet in my opinion) then for me, I would question the value of the X20S for me. And indeed some of the responses here are swaying me towards the standard X20.

To help me decide I'd appreciate it if someone could answer the following questions
  • Do the X20S gimbals have Analog or Digital Hall sensors?
  • Do the X20 gimbals have bearings, if so how many?
  • Does the X20 in black have black or silver joystick bezels?

Cheers,

Nigel
Apr 15, 2021, 06:53 AM
Registered User
Quote:
Originally Posted by Peter D Rieden
It's a matter of semantics. Hall Effect sensing is an analog phenomenon, our transmitters send the information in a digital format. Somewhere down the data chain the analog data is digitised. Whether you put the A-D conversion inside or outside the gimbal subsystem in the system architecture is just an arbitrary decision at the functional partitioning stage of the design process.

PDR
Well there is the quality of the ADC to consider, but I would imagine whether it is on the PCB as a discrete IC, part of the processor IC or on the Hall sensor IC they would all at the low budget end.

The main thing is the interconnect. The gimbal will have loose flying leads that are plugged into connectors on the PCB. PCB tracks will then take the signal from the connector to the ADC IC. It is that interconnect where interference will be picked up.

Cheers,

Nigel
Apr 15, 2021, 06:56 AM
Registered User
Quote:
Originally Posted by coreless
Thanks, that is useful. Also note that in that post Jet_Flyer is suggesting that Digital Hall sensors are better than Analog Hall sensors

Cheers,

Nigel
Apr 15, 2021, 07:18 AM
Registered User
Quote:
Originally Posted by pumuklis
Thanks to everyone for the help!
I replaced the original SD card in the X10 and reinstalled ETHOS.
IT WORKS!
I was also able to boot the R9 mini Rx.
I set up the mini quad for ETHOS and I have already flown with it in the apartment.
ETHOS still has some shortcomings on the old X10, but I hope they will be fixed.
Plain English: you don't need a X10 Express or X10S Express or the ISRM upgrade, you don't need a new radio, if you have XJT or R9M external modules you can enjoy EthOS right now on your old X10 or X10S ACCST. And you can switch back and forth between OpenTX and EthOS. You can enjoy EthOS even on Jumper T16....


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
New Product Frsky Horus X11 and Tandem receivers skorp_jon Radios 165 Apr 13, 2021 10:45 PM
New Product New Syma X20 Review Bill M - RC Micro Multirotor Drones 0 May 05, 2017 01:16 PM
News New! Ares Ethos QX 130 Jason Cole Multirotor Drone Talk 654 Sep 16, 2016 02:28 PM
Discussion New owner of a Ares ETHOS QX75 (I THINK??) Waterbound Mini Multirotor Drones 7 Apr 04, 2014 01:32 PM