|
|
Post by carbon on Feb 22, 2016 7:03:15 GMT
Looks like everybody with running projects is busy :-)
For me too..
I have made great progress.. Did some motion with the configuration tool and the simulator tool.. no real sims as I am not ready for that..
I am running in 1 particular issue.. And I don't know how to resolve.. moving the platform in the upward direction all is super smooth.. but going down I have a lot of jittering, like it is going down too fast and gets caught by the pot and stops and moves again. Or is it because there isn't enough weight on it?! Going up is not a problem , but down is a rumble!
Well I hope somebody knows how to resolve that particular issue!
One other question is how to limit the range of the lever.. is that the sensor range?
I believe at 50% is 90 degrees up and down from the horizontal position.. if I set it to 35 will it decrease ?
Like this red range I have in a simple drawing
 Where black would be 50% an red what I want :-)
I guess that no other software can overrule this range!!
Here is my current standing:
 A few more fixtures needed and then I can start building up the frame
Are you going to fit limit switches, or is over-driving not an issue with your design? If you have a sensor failure (it does happen) then any software imposed limits will be useless. |
|
|
|
Post by mhensen on Feb 22, 2016 8:07:41 GMT
I was thinking about it and over-driving is a bit of a problem, due to the upper bracing, so Yes I will be putting in limit switches.. Better safe then sorry!!
Here are my switch 'triggers' .. 135degrees is what I intend to use on the lever. As with this when the lever is at it's highest point the lever is in line with the rod. I don't think you gain much if you let the lever go past the point when lever and rod are lined up to the center of the axle. I could have gone for single' finger and then add 2 on the axle on for the low side and 1 for the high side. But this is what my design can stand best.. (I think :-), otherwise I cut one 'finger' off and reprint 6 new single 'fingered' triggers :-) ) I will rotate it just a bit so I will not trigger it when in normal duty.. but use it as a safeguard!!
The micro lever switches will come in by tomorrow so then I will design the casings for them and print them and probably wire it up slowly.. These switches will be at the outer circle of the gearbox.. attached by an M8.. will show some picture when done.. But in this the wiring is at a fixed point and the fingers should trigger the micro levers..
|
|
|
|
Post by mhensen on Feb 22, 2016 20:57:02 GMT
Ok I've build up the main platform.. and this is quite a load, I know..
But I am hitting a terrible issue now.. Well the issue was there all along but now it is unbearable..
It could be that I am hitting the limits of the star configuration of the motors and really need to step over to Delta, too bad my electrician has come around yet.. (next appointment is Wednesday!)
But for now I am still in Star configuration.. Moving up the platform works, just barely .. I see all individual motors working to get to position.. so that could be the limit of the Star config.. But going down is a whole different story.. Not sure where to search that one.. Brake resistor/power?!?! I am lost in this one..
when going down I see the levers bouncing around... ( I do this at really slow speed as I don't dear to do it fast!!!!!) and this makes scary rumbling noise and it feels like all is faling apart!!! But it is more the bouncing of the levers you hear!
I don't dare to go on that!!
PID I have had all around.. ranging from P9 up to P40.. with I and D at 0 The Break resistor is at 50% or could it be the wrong one?! they are Type: PWR-R200 100 R/J 200 Watt, 100 Ohm
I am lost in this.. hoped that moving down would be smooth..as the weight should help.. or is it something else..
Is there a setting in the VFD I've missed ?! or might this be the problem that paulg100 is referring too..
So if there is anybody that can tell me what to search for ?? Thanks in advance for that!
|
|
|
|
Post by paulg100 on Feb 22, 2016 21:13:39 GMT
whoa that is a new one on me i cant think what that might be. My motors run smooth when in motion so not the jitter i have exactly but if there is noise on your 10v analog outputs then that may cause all sorts of odd behavior.
Break resistor is for stopping the motor quicker so it should not be that, if you want to be sure you can disable it with the parameter on vfd to test.
what is online/offline motor speed on amc? when i had my offline speed set to low it seemed like there was not enough power to re-position the platform.
how were the motors running without the platform mounted, all smooth?
|
|
|
|
Post by mhensen on Feb 22, 2016 21:25:25 GMT
Yep... without platform it was smooth.. but as soon as I mounted the upper platform, even without the rig on top if was terrible on the way down.. and that was just 15kg added..
I have the offline speed at 1 as I don't dare to go any faster.. online was at 50 but I changed that to 2 also as I am not sure which one is used in the configurator tool.. but if it is online I had it on 50 and slower!
Could it be the Star vs Delta configuration?? it is just 50kg what is on top now.. !?! so that isn't much.. I don't dare to go on it.. I will first install my limit switches and kill switch as it is scary as hell :-)
I am more thinking in the direction of PID settings but I have no clue on what the values should be, could be much higher then the 25 as I am in Star config.. Or am I too much focused on star/delta?!
But any suggestion that could bring me in the right direction is welcome!!!
I will try and switch off the brake resistor tomorrow.. now it too late to make that amount of noise :-)
Hmm.. I could also try to do autotune again without rotation?? I did autotune without load on it!!!
|
|
|
|
Post by paulg100 on Feb 22, 2016 21:40:17 GMT
maybe im misunderstanding, your offline motor speed is at 1% ? if thats right it needs to be much higher, i had issues even at 15%, try offline at 30% and see if that makes any difference.
My platform was smooth even with just the primary top mount on.
Before using fabi's autocal for PID i only had P set, I and D were 0. ive had p ranges upto 70+, you will know as soon as its to high as you will get overshoot or oscillation.
if you have tried ranges between 9 and 40 it sounds unlikely the problem from my limited experience.
EDIT: I did autotune also without load, before i did autotune one motor was horribly jerky but the autotune fixed that. You have the parameter for use autotune data set?
|
|
|
|
Post by mhensen on Feb 22, 2016 21:58:13 GMT
Hmm.. ok .. Yes, I have set the VFD to use the autotune data.. I'll recheck tomorrow, you make me doubting :-)
Then I will slowly increase the on- and off-line speeds in steps of 5% to see if it resolves nor gets any better..
Edit: Just checked and H002 is on 2 so use autotune data! But noticed that B90 was on 0.5 while according to the documentation it should be on 19% so I.ve changed B090 to 1.9 as I thought it would go to 1, but goes to 10 :-) What is your F002 and F003 ?? I am at 0.5, perhaps this could be lower too...
|
|
|
|
Post by paulg100 on Feb 22, 2016 22:41:09 GMT
yep 1.9 is 19%, ive set mine at 90% (9), but i dont see how your breaking speed will cause this jerking movement during travel, i wouldn't worry about the brake resistor at the moment.
yes my f002 and f003 (acl/dcl) is 0.5 i think. again ive tried various settings with that and didnt cause the problem you have but if you want to be sure reduce it to maybe 1.0 or 2.0 to see.
|
|
|
|
Post by carbon on Feb 23, 2016 9:48:09 GMT
If I didn't know any better it sounds like a slipping or binding pot - but looking back through the thread your 3D printed arm looks quite sturdy - although if this arm can somehow flex or if the nut on the pot is loose or its shaft is binding it could be causing an issue just as you have described (happened once to me). I would disconnect the arm from the frame, remove the top wooden cover so it can free wheel if required and trying just using a weight or even your foot on the arm as a load and work from there with different motor and power settings. Also watch the pot and see if it binds/slips. Do all the motors + inverters have the same fault? In this configuration you could even disconnect the pot from the motor shaft and freewheel the motor in different direction to see how it performs.
|
|
|
|
Post by mhensen on Feb 23, 2016 13:47:39 GMT
I have tried several things but still no luck.. the lever arm starts to vibrate on the way down.. on the way up there is no problem..
had on- and off-line speed at 50% so that would be 10+ rpm, Have set the start en stop ramp from 0.1 up to 1.0, moved PID from P15 to P55 ..still no smooth descent ..
Slipping and binding isn't the issue.. The pots are firmly attached and if I loosen the nut I can turn the pot and see the arms move.. no force is needed and just a tiny bit of loosening the nut..
It is just that the whole frame start like a wild rodeo bull when moving down.. and it is on all motors.. not just one..
Worst case I will remove the cover and disconnect an arm, and make damned sure the rest of the VFD's aren't powered, and do some freewheeling rotation..
tronicgr : Do you have any idea from your experiences, can't be I am the first one with these kind of issues :-)
I tried everything as far as my knowledge goes, from slow to fast from low values to high values in P I=0 and D=0.. the lever moves down but bounces up again, goes down further and bounces up again, until it is eventually down at it's lowest position ..
always only in downward movements and even more when I am below the 50% holding position.. perhaps that is a detail you might want to know.. There must be a way to smooth it out.. (see and hear the video a few replies up, the ratling was also a loose bar on the racing construction, forgot to tighten it..
In one of the last runs the whole frame moved forwards like it was really jumping, that was a bit scary, I can't even push it back in it's original position, but it created more space to get around the back.. so I don't mind for now:-)
|
|
|
|
Post by tronicgr on Feb 23, 2016 21:07:13 GMT
Something is slowing down the PID loop. Check to see if you have the sensor oversample down to 1 and while testing manually with the buttons, disable the LCD menu (push down the DIP switch "1").
Other things to check is the minimum speed to be set at 1, and keep the PID low like (P 15, I 0, D 0)
What is the Max RPM rating of your gearbox?
Thanks
Thanos
|
|
|
|
Post by mhensen on Feb 23, 2016 22:47:37 GMT
Something is slowing down the PID loop. Check to see if you have the sensor oversample down to 1 and while testing manually with the buttons, disable the LCD menu (push down the DIP switch "1"). Other things to check is the minimum speed to be set at 1, and keep the PID low like (P 15, I 0, D 0) What is the Max RPM rating of your gearbox? Thanks Thanos max RPM is 23.3.. but now at 50% so it should be 11.6
I will try these settings tomorrow but I think I have played with them before, but I more thinking about that the motors are struggling.. and when going down it has both the weight and the downward rotation of the lever..I think that due to not enough torque if falls through the requested speed and thus tries to recover from that by pushing it back to the position it would be at that particular time.. the gearbox torque goes up to 200nm but that would be on a delta config of the motor, not Star, so I would be on 1/3 of that 200nm I think!!! Thus gravity in combination with the current max torque is the problem?!?
Could this reasoning be valid?
Well tomorrow the electrician should swing by and that would give me 2*16 amps.. so I will then reconfigure the motors to delta.. do I need to autotune again, before I start everything up again?!
|
|
|
|
Post by mhensen on Feb 24, 2016 16:34:45 GMT
Well today, my new power lines were installed.. I have got 2 * 16A now so will run VFD 1,3,5 on one and 2,4,6 on the other.. Should be enough to carry on in Delta configuration.. I am doing a autotune, without rotation, but also install the limit switches before I can turn it on all at once.. Installing the witches is a nasty job as you need to now the rotation direction.. You don't want to mess that up, otherwise the switches have no use.. So it will take me time before I can test the new setup in it's full glory and hopefully the rumbling bucking bull type motion downward is gone..  With the 2 triggers on the axle I should have 135/140 degree of rotational freedom.. which is fine with me and the frame.. And I need a nice cover on that nasty part of the lever :-) that will be a new printjob |
|
|
|
Post by wannabeaflyer on Feb 24, 2016 18:25:08 GMT
Hi Mhensen , been lurking on this site watching the great progress you guys are making with your projects and the ongoing work getting the systems up and running, this is not a critisism just a suggestion regarding your limit switches arrangement. My project is only a 24Volt DC based 6DOF and in no way compares to what you guys are doing here , but I thought I'd pass on a safety tip I learnt along the way ,
Your switches are quite small given the size and power of your system, I may be wrong but if your going to use switches that small then I believe you will benefit from creating a Cam type switch trigger rather than the lever system your currently using...
No way is this suggestion picking holes in your work or method good buddy , the logic behind that was that, with these small switches the contact make or break gap is very small and any over travel will generally destroy the switch , in that case this may or may not stop further rotation of your actuator arm but sometimes the damaged switch state could lead to false signal which in the worse case could mean full 360 rotation ,,, you guys have seriously powerful rigs, so i'm guessing that's a condition you want to avoid. with the cam type switch trigger any slight overshot due to say momentum / Inertia would still trigger the switch but not damage it...
As I mentioned above please this is not criticizing your work just my two cents based on lessons learnt in the past .... Im not Yet an AC sim Builder But that is the dream, Im Going to be Using My AMC128USb controller to drive my 6DOF sim , Small fry compared to you guys but something transfer across no mater what size .. Keep up the mega work Mhensem I for one check in everyday to learn form you guys - |
|
|
|
Post by mhensen on Feb 24, 2016 20:28:52 GMT
Thanks @wannabeflyer for tip!! (seriously appreciate it!!! Just in case the rest might telling you differently :-) (English is not my native tongue so I might phrase are misspell things! )
No I don't mind you picking holes in my work :-) .. No problem at all, I'd rather have it then that later you might say that you already had the notion:-)
You are right that the switches are damned small, but the thing is, it is just a catch to cut off the motion in a particular direction momentarily..
With a cam type I have the feel that once hit you have to 'reset' it.. In that case you might end up in a situation that is even less preferred, with the platform in kind of weird angles.
But I might be wrong in this !!!!!
Anyway in case of a trigger, there is something wrong in either hardware or software settings.. The 'trigger finger' ring is just clamped plastic on a smooth iron axle, so in case of trigger hitting, it can slip with the rotating axle while keeping the switch pressed, just in case the deceleration is not small enough.. The switch just kills the directional output of the VFD, not the rest of the motion.. For that I have a cam type switch in the 'cockpit', that one will kill all motion.. I have been using these switches for the last 7 years in my 3d Printers as stops.. I know that the forces are completely different, but I have been abusing them heavily and yet has to see one fail..
So in the end I will, if all is correct only hitting the triggers when setting up any motion queues, in that case I will have to tone done all the queues to make sure nothing gets triggered.. otherwise it should never happen that it get's triggered! Just an extra safety barrier!
Here is a little test, it might be hard to see and I will shoot a new one during daylight .. but this is the idea.. Motion should be blocked but as soon the other direction kicks, the motion should and must continue to recover to a normal state..
|
|
