Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: turbothis on June 01, 2011, 02:18:23 AM
-
why would x or y be smooth on there own but both rapiding together makes for a very jittery movement? it rocks the mill pretty good.(5000#)
it is only 80ipm rapid
the servos are tuned well and have plenty of power 80v 40A
i have searched but cant hit the keywords to get the results i am lookig for....
thanks
-
What do you have the Kernel Speed set at ? (Config : Ports & Pins).
Tweakie.
-
100k
does it sound like the computer wont keep up with that many pulses?
-
I suggest that you try changing the Kernel Speed to 25k, restarting Mach and retuning your velocity settings for maximum reliable motor speed and save these settings.
100k is really not the best setting for performance and, depending on steps per, it is unlikely that it is ever necessary to exceed the 35k Kernel Speed.
Tweakie.
-
well, i am going 1024 on the encoder, 4-1 on the pulley ratio, and .200" screws. 80 ipm rapids
i think it was like 92k or something on the motor tuning.....
-
I would change out the encoder to about 1/4 of that. No way can you use that much resolution. 128 encoders would give you less than 1 ten thousand inch resolution .00009".
Let's be resonable, (;-0 TP
-
I would change out the encoder to about 1/4 of that. No way can you use that much resolution.
Have to disagree there, the more resolution you have the better as it should allow much better tuning on a servo system but thats just my opinion ;)
But with high count encoders and the parallel port it can be a problem as you need to run a high kernel and not many will reliably run at over 65KHz. Using an external controller such as SmoothStepper, KFlop etc does away with that issue.
See if your drives can utilise electronic gearing, personally I am not a fan of it but at least you would be able to reduce your kernel without having to change out encoders.
Hood
-
At better than 81000 PPI you are long past the point of diminished returns on resolution. Take a look at the resolution of high end commercial machines you won't find many if any that high.
Remember this machine has 20x reduction on the servo. If you don't want to change the encoders then change the 4:1 to 2:1 reduction or 1:1
(;-) TP
-
Terry I am meaning resolution between drive and motor and most industrial servos with encoders that I have seen have at least 2000 counts per rev, some, such as the MPL motors I use on my spindles, which have smart encoders, can have as much as 1million lines per rev if required (I just have them set to 4096) .
Hood
see you have edited ;) that is why I mentioned electronic gearing, gives lower resolution between Mach and drive(axis movement) but still keeps the same between encoder and drive.
-
YES I agree that electronic gearing can work BUT you still end up with micro cogging you try toget rid of with high encoder counts (;-) The gearing just uses blocks of pulses instead of one.
YEP here I run 2000 count encoders and end up at 40000 PPI and that is WAY more than is needed to have smooth motion, high resolution @ .000025" ( way more than this ole Bridgeport can use) and 100 IPM rapids.
Just a thought. (;-) TP
-
i dont do anything crazy accurate.
so it looks like i should start low on the encoder count and work my way up to the performance level i need?
my encoders are pin adjustable. they are the cui 102's
-
The way I am seeing it is if you keep the higher count encoder then its making things easier for the drive to keep things in position as it should be able to react quicker to motor movement, assuming the drives are capable, which I am sure they are. I personally dont like using electronic gearing as I normally have no gearing but because this setup has a high gearing then I dont really think cogging would be an issue.
Of course everyone will have their own views ;)
Hood
-
what is electric gearing?
i am running 320x's
-
Personally I would use electronic gearing if your drives support it. As said I am not a fan of it but with the high gearing you have it would be fine I think.
Hood
-
Electronic gearing is where the DRIVE can take 1 pulse from the controll then output a mulitipled BLOCK of pulses (Say 10x) to the motor for each input pulse.
IF you can "simply" change the resolution of the encoder, try it you may like it. Don't forget to reset the motor tuning to match.
IF not then try other approachs.
It is always about the final results rarely about how you get there, UNLESS you have restrictions in hardware. Then you are limited to approaches that use the GIZMOs you have to work with.
Just a thought, (;-) TP
-
thanks guys, i will play with it some.
-
You may also need to retune your servos if you alter the encoder count.
Hood
-
so i got my encoders down to 200 count, 4-1 belt drive, and 5tpi screws.
i can net 130ipm pretty easy and she is very much smoother on the xy jogging together. i did notice the speed says it is doing like 170ipm which i thought funny because the motor tuning is @ 130ipm. how does that work?
-
Is that when moving two axis that you see the higher rate?
Hood
-
yes
both xy are set to 130ipm and when you jog both at the same time it goes up to around 170ipm on the mach speed tach thingy. the table seems to be doing just 130 though. i can hear it running the same speed as it should.
-
That is the blended speed you are seeing, if you jog a single axis or command a diagnonal move with a G1 and a Feed of 130 it will move at 130.
Hood
-
oh i see.
i just thought the max being at 130 it would do that diagonally too.
thanks
-
The max is 130 per axis but because its travelling diagonally it is moving a further distance so the blended feedrate will be higher.
Hood
-
gotcha
-
But as said earlier if commanding a feedrate on a diagonal then that would be what you get, ie X and Y would move slower so that the blended feedrate was correct.
Hood
-
ya, that makes sense.
i got a smoothstepper ordered and will see how awsome it can make my machine. >:D