FlyPT Mover

So I finally get Flypt to work on my 6dof. I wrote myself a small program that will copy and convert data (from double to float 32bit) from MMF to Serial port. I then setup the Flypt to output to MMF....not too pretty, but it work. Hopefully the next software release will include the option of out put Just Pose -> Rig pose in float 32bit

I have to say Flypt is by far the best mover program I have seen. (Thanks you for your great work). But the learning curve for setting it up properly is not a easy task, need plenty of reading and study...especially the filter part, very powerful but take lots of learning. My ride now is running okay, but will still need some tuning to perfect.

The source recording and viewer graphic is very useful to compare the input vs output. I am thinking to add a data source of 6dof sensor (3-axis accelerometer, 3-axis gyroscope) to measure the actual movement of platform. This way, I will be able to compare the actual motion feedback vs game data.

Tony

I am having some issue to setup the heave in Flypt, can someone help me out.

my 6d0f platform has heave scale from 0 to 200 (mm), I use vertical accel + gravity, which is 9.8m/s2 when vehicle is in still position. My filter of heave is MAXCROP(MINCROP(VALUE;0);200.00), this is to make sure the output value will be within my range.

The problem I have is that when vehicle jump and landed, the output duration of that peak value is too short and so basically my platform don't have time to react. When vehicle landing, the value v.accel+gravity is reduce from peak value to 9.8. during this, driver shall continue to feel the extra g-force. To simulate that, the heave shall continue positive movement until the v.accel+gravity reach 9.8m/s2 again.

Is there anyway we can have current_heave = previous_heave + vertical_accel * constant

This way, as long as the vertical_accel is positive, the heave will continue raise to feel positive G
and if the vertical_accel is negative (during air), the heave will continue lower to feel negative G

I did above with my program (only work with dirt rally), and it work pretty good. I am wondering how to achieve something similar with filter with Flypt.


Thanks
Tony

@itonylee,
Give me some hours.
I will say something later.
Familly obligations...
Sorry

Definitely, family first!

Question to everyone playing GTA V. There's a genuine SOURCE module saying it needs a HOOK. But the help button is a dead link, there's no GTA V chapter in the online help.
Is the hook Knaufinator's ScriptHookVDotNet3 plugin I used when I played GTA V with SimTools?
One more question: I haven't played GTA for quite a while and must install it from scratch. I own both versions, the EPIC store's and Steam's. With the WMR G2 and the R.E.A.L. VR plugin needed, which version would you recommend?

Mover 3.4.2

I may have mentioned this earlier but has there been some major change to the Spike filters?

I find until here we dont filter at 0 feels like its at 3-4 and its only for huge spikes. The other two inputs does work.

Could this have something to do with the scaling and gain of the rig, it just dont feel right.

the rig is set to 1:1 using a gauge on the rig so 10 degree is correct. i Find i need 5.000 (50 degree) in pitch/roll there if that matters at all

I did a test with SPIKE 0:2:3 and it felt the same as 0:20:10.

Edit.

It seems that Spike filters now need 100 and 200 instead of 1 - 2 which is so far out of what i was using before that i just didnt test it.

Hello everyone,
I got some issues with the serial output.

I found a lot of lags in the serial packets, there some dangerous burst at 115.200 bps. There is a sort of expansion of the curve over time then the platform freeze for about 15 ms and restart very fast recovering the lost packages all in one time. It happens principally when the geometry gets out of range/singularity. can I remove these strange lags?
In the two pictures with the blue pulses, is possible to see the delay just off the serial, tested with an oscilloscope. These show that when the singularity, or an always precise position, is reached, stop to send position for a while (about 15-30ms), then send all the buffered information all in one time so that a burst happens.

In the picture with the purple numerical output, in the black rectangle of the serial log, we see that at a certain point the actuator goes into a singularity, he stops changing his position, but the others keep moving. shortly afterward, going forward, downwards, we see that all the pistons stop.
In singularity, there is only one, or maybe two
but the worse is that, in the red box, you see that when they start to move again, one actuator continues from the point where it stopped, but the others jump by thousands of impulses ...

In case, I could send some video of the rig
I don't know if I'm been able to explain correctly the issue, I hope that someone could help me

Hi,

I'm going to look at it tomorrow. This and other problems and requests.
Hope to get a new release this week.

Need to update the pointers in Mover.
That's the problem of using hooks. Any update in the game and we loose them.
One of the reasons I created a custom hook source, that allows users to put updated hooks.

This takes time to do, and right now I'm short on that. But I will try to update them for the next release.

Really sorry...

I promise I will look at it better tomorrow. Got an almost free day from work.

I chaged something in the spikes....
But with so many updates I'm now lost. I have to look better to the code. But I know it was wrong before.
Tomorrow, I promise.

Appreciate that. Ive noticed the spike filters do crash from time to time too and just leaves the rig completely stuck when using higher values.

Another issue that may be related is when i was using FFB signal to the SFX100 (trough Cubexx ardun Mega sketch if that matters) it could crash completely and giving heavy noise/getting stuck. happened 2-3 times within a short amount of time but i havent tested that much as i have slam the rig down.

@Marco Annunziata,

The problem, seems related to some mix in the ui and calculation thread.
Looking for that hole right now.

Ive mentioned it a few posts back but if you could fix the condor source for directs so they save it and show correct it would be epic for next release. Im about to start Sail flying and re enabling 6 directs every time quickly gets old. It wont save.

I think it's something to do with the directs. I have to assign mine every time, as they do not save in the output.

Its only Condor.

Hi chaps.

I am trying to make a noise from value for the Canopy in DCS. The problem is that the value goes from 0.0-1.0 as the canopy opens but i need it to be 100% all the time and only when the canopy is moving. I can use a L spline but i cant put a high pass on it so if it sits in the middle and not moving i will get the noise. Ive tried different combinations with HP but it does not work well since the value starts low and goes high.

Things seems to suffer from not working when the values reduces.

As far as i can tell this is also a problem with transdusers?

Anyone got this nailed down ?

Edit.

Seems to work with straight HP then add another source for HP with gain -1.0000 for the close.

Would be nice if it worked on both up and down so you dont need to make two. Unless that causes a lot of issues.

I have added my profile if anyone wish to see how its done. I have added both a rig noise and transducer.

Sadly there is some lag in between things but you dont really notice when opening and closing fully

Make sure you update for your rig if you use it.

3.4.2

The rig will jerk as its got lots of noise on and go up to about where it should but just spike all over the place. Feels like it has something to do with the spike filters.

It's happened a number of times.

It's everything for me using 3.4.2 Thanos output, my Direct never saves. However it works fine with 2.9.9.

I have to admit i didnt check to closely as it works with DCS . While it does not say Air/Ground for my DCS it does save both. Xplane does however fail to save and will always default to the top one.

My apologies.