Force Feedback Joystick Coming soon - Page 11 - ED Forums
 


Notices

Reply
 
Thread Tools Display Modes
Old 08-22-2019, 09:09 PM   #101
Berniyh
Member
 
Join Date: Dec 2015
Posts: 104
Default

Quote:
Originally Posted by Drakoz View Post
Advanced versions of PID loops have factors that explicitly counteract gravity while dealing with a mass moving in angular directions (just like a joystick), and they do this while maintaining accuracy within 0.0005 inch (0.001mm) or better. There is no reason why a 2 axis joystick can't do the same.
Unfortunately, there is.

Keep in mind that CNC machines (and similar tools) work in a rather predefined environment with predictable movements (or at least predictable movement directions).
This does not apply to FFB in a simulation.
The problem here is not that it's not possible to compensate for gravity (it definitely is).
The problem is that it's very hard to tell if a movement/force is induced due to gravity or due to user input.
With detailed profiling you might be able to achieve it for a specific base/stick setup.
But even then users might complain that the stick feels very weird when they move the stick in a certain way due to the gravity-compensation kicking in and you would notice that even if it was just a slight click for a short time.

tbh, the much easier solution would be to put a proximity (or contact) sensor into the stick detecting if there is a hand on it or not, so I really doubt that any manufacturer of an FFB stick would put that amount of effort into gravity compensation.
Especially since we're talking about an issue that isn't that much of an issue in the first place (since taking the hand of the stick isn't something you're going to do on a regular basis, especially not in a helicopter which are almost exclusively the aircrafts using force trim).
Berniyh is offline   Reply With Quote
Old 08-23-2019, 12:48 PM   #102
Drakoz
Member
 
Join Date: Sep 2014
Location: San Jose, CA
Posts: 199
Default

Berniyh, I am confused at your response on this. You own a Brunner don't you? Does it fall forward when you let it go (assuming it has no hand sensor or you have disabled it)? Perhaps only if you have the stick spring force set really low, but even then, if it falls, it should be really minimal. I am curious to know.

What if you smack it (again with the hand sensor disabled)? Does it go into oscillation. A G940 (which will oscillate itself to pieces) or MS FFB2 (which will usually damp itself out in a couple seconds), both have horribly unacceptable oscillation because there is no PID loop involved (just the P part of it). Their solution was to add hand sensors (or maybe the lawyers required it), but that is also unacceptable to have the stick go limp just because you let it go. But a Brunner or Gauss should not have these problems because they should have good motors, no backlash in the gimbal/drive system, and because they are using the kind of servo loops that I am talking about. They would have to. And if they don't, they aren't worth the price.

I think I hear you saying you are concerned (rightly so) that trying to solve these issues might add some aberrant force feeling (a bad feel in the stick - a bad "force profile"). Is that the concern? If so, don't be concerned. I don't know how I can say it better without writing another book. Again, I do this kind of thing for a living. A basic PID loop can achieve what I am talking about without any weird feeling or force profile. In fact a PID loop isn't a force profile - there is no profiling involved (i.e. custom curved force profiles). Just equations that achieve a specific goal (spring force which I guess is the force profile, but this is just a linear F = P * X force, damping, and holding accuracy), but are limited in their implementation so the only thing the user feels is smooth spring force application and maybe some intentional hydraulic or friction feel (which is what simFFB does).

Compare that to a G940, where we feel horrible bounces in the force as the software fights to deal with its loose gimbal and gears, and sometimes even fights us (and the gear mesh and motor pole feeling is just horrible too). But that is due to slop, horrible motors, and bad software (which free41's patches helped improve a lot). A PID loop would not fix a G940 because the bad motors and gimbal slop are too much to fix in software alone. But on a good mechanical design, the end result of a proper PID loop is improved feel, not a ruined feel. The Gauss and Brunner should have that improved design.

Last edited by Drakoz; 08-23-2019 at 12:52 PM.
Drakoz is offline   Reply With Quote
Old 08-23-2019, 04:35 PM   #103
Berniyh
Member
 
Join Date: Dec 2015
Posts: 104
Default

Quote:
Originally Posted by Drakoz View Post
Berniyh, I am confused at your response on this. You own a Brunner don't you? Does it fall forward when you let it go (assuming it has no hand sensor or you have disabled it)? Perhaps only if you have the stick spring force set really low, but even then, if it falls, it should be really minimal. I am curious to know.
Yes I have a Brunner and no it doesn't have a hand sensor.


afaik there is no algorithm implemented in the Brunner to prevent it falling of, so it does of course.
how much depends on the force profile, which is expected.
The problem here is that if you trim the stick and then put your hand off the stick even small movements will cause the aircraft to run out of a stable position (how quickly depends on various parameters including the aircraft itself).
Unless there is a mechanism specifically preventing this (and there isn't) it will always happen, it doesn't depend on the torque capability of the base (which was the original question I responded to).


In the end if you're happy to use rather pronounced forces you can definitely set up the Brunner (also the Gauss) such that the fall off will be minimal, but you have to consider higher moving forces. At least for a heavier stick.


Personally I'm using the Brunner with the Warthog stick and I'm also using rather low forces since I'm rather after sensitivity and precision.
In my case the stick will fall off quite a bit, but that's ok for me since I just keep my hand on it. The trim for me is there to relax the hand and to have more precise control around the trimmed position (meaning that the forces on my stick are tuned for precision rather than feedback).


Quote:
What if you smack it (again with the hand sensor disabled)? Does it go into oscillation. A G940 (which will oscillate itself to pieces) or MS FFB2 (which will usually damp itself out in a couple seconds), both have horribly unacceptable oscillation because there is no PID loop involved (just the P part of it). Their solution was to add hand sensors (or maybe the lawyers required it), but that is also unacceptable to have the stick go limp just because you let it go. But a Brunner or Gauss should not have these problems because they should have good motors, no backlash in the gimbal/drive system, and because they are using the kind of servo loops that I am talking about. They would have to. And if they don't, they aren't worth the price.
That's a common problem of feedback devices.
You can definitely setup the Brunner (and also the Gauss) such that it will happen.
I do also own an OSW (Simucube 1 with Lenze servo) and you can get oscillations there as well.
This is not about the motors, instead it's a rather basic issue with such feedback devices. It's also not about price but setup and software, basically how you interpret the signal.
If you just pass on the FFB signal to the controller without any kind of filtering, every feedback device will start oscillating at some point.

To better understand it, compare it to the situation where you put a microphone in front of a speaker that puts out the sound the microphone records.
I'm pretty sure you know what would happen, you're starting to hear very loud noises.
This is very similar, because what you created is a feedback loop.
The solution to this problem obviously is to separate the microphone and the speaker such that the microphone doesn't record the speaker anymore.
Unfortunately such a solution is not possible in case of our feedback devices because input and output are bound to each other by definition.
To prevent such things to happen you have to add filtering and post-processing to your signal.
e.g. a very common way is to add damping to your axis. If there is enough damping then there will be no oscillation.
The problem here is that for obvious reasons damping also kills feedback.
So in the end there is a trade-off between direct and defined feedback and preventing oscillations.
There are advanced filters that can result in a better trade-off, but in the end you will always wash out some of the feedback as well if you reduce the oscillations.
In the end regarding the oscillations the same thing applies as to the fall-off. It's only a problem if you take the hand off the stick and since that isn't something I'd recommend, I'd say it's not much of a big deal.
(Still there are big discussions about this topic on the sim racing forums. tbh, no idea why everybody seems to be so eager to take off the hands in a race car at 300 kph …)

Quote:
I think I hear you saying you are concerned (rightly so) that trying to solve these issues might add some aberrant force feeling (a bad feel in the stick - a bad "force profile"). Is that the concern? If so, don't be concerned. I don't know how I can say it better without writing another book. Again, I do this kind of thing for a living. A basic PID loop can achieve what I am talking about without any weird feeling or force profile. In fact a PID loop isn't a force profile - there is no profiling involved (i.e. custom curved force profiles). Just equations that achieve a specific goal (spring force which I guess is the force profile, but this is just a linear F = P * X force, damping, and holding accuracy), but are limited in their implementation so the only thing the user feels is smooth spring force application and maybe some intentional hydraulic or friction feel (which is what simFFB does).
ok nice, but I would have to see it in action and so far nobody implemented anything like that as far as I know.
And I'm pretty sure you won't find it in this base either.


Edit: also sorry if I told you things with the above that you already know, just wanted to set that straight regarding the oscillations.
afaik there is no consumer FFB device on the market that does not start to oscillate unless you apply some kind of filtering or damping.
Even the most high-end stuff like the Simucube, the Fanatec DD and the Bodnar wheel do that and some of them have been in the business for years and in case of Granite Devices (Simucube) they've also got years of experience with servo drives and feedback devices in industrial applications.

Last edited by Berniyh; 08-23-2019 at 04:39 PM.
Berniyh is offline   Reply With Quote
Old 08-23-2019, 05:22 PM   #104
FoxHoundcn
Junior Member
 
Join Date: Oct 2015
Posts: 97
Default

We are already discussing this issue and intend to provide a calibration to eliminate the problem of too heavy joysticks. This calibration is similar to trimming and is handled with a default trim offset. We will continue to try and bring better solutions.
FoxHoundcn is offline   Reply With Quote
Old 08-23-2019, 06:41 PM   #105
jojo
Veteran
 
jojo's Avatar
 
Join Date: Oct 2008
Location: France
Posts: 3,553
Default

Where does this Gauss F-15C comes from ?
I can't find anything about it...
__________________
Mirage fanatic !
I7 7700K/ MSI GTX 1080Ti Gaming X/ RAM 32 Go 2400 Hz/ SSD Samsung 850 EVO/ Saitek X-55 + MFG Crosswind + Rift S
Flickr gallery:
https://www.flickr.com/gp/71068385@N02/728Hbi
jojo is offline   Reply With Quote
Old 08-23-2019, 08:24 PM   #106
FoxTwo
Member
 
Join Date: Mar 2019
Posts: 125
Default

Quote:
Originally Posted by jojo View Post
Where does this Gauss F-15C comes from ?
I can't find anything about it...

It comes from them. They're Gauss.
FoxTwo is offline   Reply With Quote
Old 08-24-2019, 05:41 AM   #107
crazysundog
Member
 
crazysundog's Avatar
 
Join Date: Apr 2008
Posts: 675
Default

I am very interested in seeing more of this product;
it is what I've been searching for, have used a MsFFB stick for 11 years now.

Great looking product!
crazysundog is offline   Reply With Quote
Old 08-26-2019, 08:57 AM   #108
amazingme
Member
 
Join Date: Nov 2014
Location: Bucharest
Posts: 834
Default

FINALLY!!!!!
__________________
Specs:
Spoiler:
Asus Z97 PRO Gamer, i7 4790K@4.6GHz, 4x8GB Kingston @2400MHz 11-13-14-32, Titan X, Creative X-Fi, 128+2x250GB SSDs, VPC T50 Throttle + G940, MFG Crosswinds, TrackIR 5 w/ pro clip, JetSeat, Win10 Pro 64-bit, Oculus Rift, 27"@1920x1080

Settings:
Spoiler:
2.1.x - Textures:High Terrain:High Civ.Traffic:Off Water:High VisRan:Low Heatblur:High Shadows:High Res:1920x1080 RoC:1024 MSAA:4x AF:16x HDR:OFF DefS: ON GCI: ON DoF:Off Lens: OFF C/G:390m Trees:1500m R:max Gamma: 1.5
amazingme is offline   Reply With Quote
Old 09-04-2019, 04:09 AM   #109
FoxHoundcn
Junior Member
 
Join Date: Oct 2015
Posts: 97
Default

Attached Thumbnails
Click image for larger version

Name:	微信图片_20190904120727.jpg
Views:	451
Size:	111.7 KB
ID:	216688  
FoxHoundcn is offline   Reply With Quote
Old 09-04-2019, 10:32 AM   #110
Supmua
Member
 
Join Date: Feb 2018
Location: USA
Posts: 628
Default

The line up photo looks pretty awesome. You can tell which grips are metal from just looking at the shine.
__________________
PC: 9900K/2080Ti, 32GB RAM.
Joystick bases: TMWx2, VPC WarBRD, VKB GFII, FSSB R3L
Joystick grips: TMW, TM F/A-18, F-16SGRH, Kosmosima LH, MCG, MCG Pro, MongoosT50-CM2
Throttles: TMWx2
VR: HTC Vive/Pro/Oculus Rift/Valve Index.
Supmua is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT. The time now is 03:49 AM. vBulletin Skin by ForumMonkeys. Powered by vBulletin®.
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.