|
Post by hexpod on Nov 1, 2018 19:42:12 GMT
UPDATE the problem maybe is the motors are breaking to bad. they running out to long and it takes to much time to change direction. the direction sensor gets out of range and the motor is turning on in direction to get position back. hmmmm now i changed the parameter dec in the inverter from 3.0 to 0.1 it seems that the motor now is not running out much and changes the direction much better. i moved the slider for about 4min from left to right an nothing happend no 360turn. i will test it with all inverters but for now it looks better. can someone tell me if this could be the problem? View AttachmentIn heXpod you can run the heave loop feature with settable speed, amplitude and stop time while tweaking your PID and inverter settings. You should be able to achieve fast and accurate arm displacements similar to that: vimeo.com/263733345
|
|
|
Post by hexpod on Nov 2, 2018 0:55:22 GMT
UPDATE the problem maybe is the motors are breaking to bad. they running out to long and it takes to much time to change direction. the direction sensor gets out of range and the motor is turning on in direction to get position back. hmmmm now i changed the parameter dec in the inverter from 3.0 to 0.1 it seems that the motor now is not running out much and changes the direction much better. i moved the slider for about 4min from left to right an nothing happend no 360turn. i will test it with all inverters but for now it looks better. can someone tell me if this could be the problem? View AttachmentOnly for your information, I've set my accel/deccel in my servo to 16ms. With values higher than 80ms my rig was unable to transmit fast vibrations. Maybe you could experiment with lower accel/deccel values than 100ms Take your time as a good sweet spot between acce/deccel and PID can be a time consuming process.
|
|
|
Post by onhill on Nov 2, 2018 10:26:54 GMT
thanks for the info hexpod your servo motors are damn fast wow very cool i found a way in the inverter manual to get acc and dec to the lowest to 0,05s (50ms) full motion test youtu.be/yBJUuV4bh5E
do you think the test looks good?
|
|
|
Post by onhill on Nov 2, 2018 10:34:22 GMT
another question would you possibly have a dirt rally 6dof profile for simtools? I have already installed the hexpod + dirt rally plugin after the instructions of thanos
|
|
|
Post by hexpod on Nov 2, 2018 11:33:04 GMT
thanks for the info hexpod your servo motors are damn fast wow very cool i found a way in the inverter manual to get acc and dec to the lowest to 0,05s (50ms) full motion test youtu.be/yBJUuV4bh5E
do you think the test looks good? I see your platform is "swimming" a bit. It could maybe come from the P being to low. What is your P value at the moment? I would try to tune the PID in the AMC in a way so the assigned position could maybe reached faster.
|
|
|
Post by onhill on Nov 2, 2018 12:10:52 GMT
P is 60 at the moment
I and d are 0
|
|
|
Post by hexpod on Nov 2, 2018 13:32:05 GMT
P is 60 at the moment
I and d are 0
It is hard to see in the vid. It can be that your P is to high and you have some overshoots.
Could you please make a test with a simple heave loop and 1000ms wait time?
Does the arms reach accurately their position?
|
|
|
Post by onhill on Nov 2, 2018 16:49:27 GMT
hmm yes looks like overshoots here is the video heave + - 60mm test sequence 2522 wait time 1000ms
motor speed is 1435rpm gearbox ratio 51 gearbox speed 27rpm
|
|
|
Post by hexpod on Nov 2, 2018 17:54:39 GMT
hmm yes looks like overshoots here is the video heave + - 60mm test sequence 2522 wait time 1000ms
motor speed is 1435rpm gearbox ratio 51 gearbox speed 27rpm
increase the wait time to 2500ms and try to find a P value maybe between 25 and 40.
all 6 arms have to stop each time exactly at the same position
|
|
|
Post by onhill on Nov 2, 2018 18:46:22 GMT
OK I did it's set to 10 at the moment
The motors are now stoping much better
But it's way to slow or?
|
|
|
Post by hexpod on Nov 2, 2018 18:54:14 GMT
OK I did it's set to 10 at the moment
The motors are now stoping much better
But it's way to slow or?
Yes that's ridiculously slow. Maybe you could compensate the P overshoot with derivative value. It's a complex topic. Such a try error process can take long time.
|
|
|
Post by onhill on Nov 2, 2018 21:02:52 GMT
ok that's right you really need a lot of time. is it even possible to adjust the values so that everything fits perfectly? You use servo motors which are already much more accurate is it even possible with my motors?
it would be best to adjust the values when the complete seat is mounted or? best to use sandbags for weight?
|
|
|
Post by onhill on Nov 3, 2018 12:57:46 GMT
Yesterday I tried some time to tune the pid but I did not make good progress Is there a good way to set the pid? Or what should you pay attention. I will mount the seat on the platform in the coming days. And wait until my 12 emergency switches are delivered. I'm really afraid that the platform will do whatever it wants and destroys itself or hurt someone
|
|
|
Post by hexpod on Nov 3, 2018 17:50:43 GMT
Yesterday I tried some time to tune the pid but I did not make good progress Is there a good way to set the pid? Or what should you pay attention. I will mount the seat on the platform in the coming days. And wait until my 12 emergency switches are delivered. I'm really afraid that the platform will do whatever it wants and destroys itself or hurt someone
Although your limit switches can give you a security feeling, they are not the right way to solve your issue. Your arms have to go reliably to the assigned position. This is a job of your inverter and controller PID settings. Back then the guys were using a tool for that purpose. As far as you stay on the AMC 8bit output you can play around with this tool. More about it here: motionsim.freeforums.net/thread/68/fabis-thanos-board-configurator-beta?page=6Good luck
|
|
|
Post by hexpod on Nov 3, 2018 17:59:44 GMT
Yesterday I tried some time to tune the pid but I did not make good progress Is there a good way to set the pid? Or what should you pay attention. I will mount the seat on the platform in the coming days. And wait until my 12 emergency switches are delivered. I'm really afraid that the platform will do whatever it wants and destroys itself or hurt someone
Please show a short vid with a resonable P value (like 25 for ex.) I and D at 0. Choose a loop with high amplitude (like +/- 120 - 150), set the full speed with 3s wait time. Thanx
|
|