[gfacer2]

Hi All,

I've just gotten a new chinese cnc router and I'm working out the kinks. The main thing that I am working on is some chatter like lines on the edges of my circles and diagonals. I've detailed it elsewhere but bottom line is I'd like to take a stab at using the autotune function and I think I can sort it out for the x axis drive (with all the skill a political science degree offers) but not sure after extensive searching how you tackle the Y when it has a slaved 'a' (I think).

What do I need to do to tune each of the Y drives? Disconnect one and run the autotune one at a time or disconnect and run autotune on one and then manually type the same settings into the other? Or might the settings for the X be appropriate to the Y as they are on the same drive system (though the weight would differ significantly). Not sure yet which is Y and which would be A...

The motors are SGM7G-09AF C61 and the sevo packs are SGD7S-7R6A00A (002 on the box but not on the pack)

My controller is a syntec 6md which has very very limited documentation (and for now I'm assuming I've gone further than factory 'support' will go) and the parameters seem to be controlling the drives with a/b which I believe is step and direction. I've tried parameters related to x axis loop gain (and Y but it threw up a regenerative positioning error (which as it is a translation - might not make any sense) so I reverted to factory setting) and smoothing and some accel/deccel response times but no change noticeable. I should note that I assumed going in that I'd be on my own for stuff like this so that's why I am looking at a DIY approach rather than complaining to the factory in China.

The installation manual for another syntec controller has been of some help in the parameters but the way it describes the setting up of the Mitsubishi drives is a very poor translation and not a direct comparison at any rate. Nor does it say anything about slaved drives (nor do any Yaskawa videos etc). I've also looked at a centroid tech bulletin (as I had centroid and thought it might point me in the right direction ) but they don't go into slaved axes either (as mostly they do mills which wouldn't have them) though I can touch base with my old centroid guru and see if he has any paid advice on this (but again its apples to mandarins in this case).

Also on the off chance anyone that has done this with a syntec 6m series, any pointers would be greatly appreciated though I plan to attempt the x this evening so fingers crossed. I'll post again if I can or can't figure out the X.

[mactec54]

gfacer2

You can go to the Yaskawa web site and download the Drive setup software, you will be able to see what the drives are doing instead of guessing, which you can't really do with these Drives, you will need the cable that goes from the Drive to your PC as well, autotune may not help you very much if any parameters are to far off, this is just a Parameter to turn on to activate the auto tune

[gfacer2]

Yeah I've got the sigmawin+ installed - I have the cable and found the drive OK. Backed up to start. Entered the encoder pulse per rev from the Syntec parameters on the top of the electronic gear and the x axis pitch on the bottom in basic length units (BLu) which are .001mm (so 1048576/27780). I have tried the autotune twice now but getting a failed result - it suggests changing the mode (not sure what mode) and adjusting the position complete setting from 7 reference units to a higher number, so I plan to bump it up to 10 to start and go from there.

The tuning itself seems to be working similar to the autotune I had on a centroid servo system (though a little less violent than that one was).

Still not sure how to do the Y though as stated. Anyhow - if you see any glaring errors in the above let me know and if you have any suggestions on the slaved Y please advise. I know you know your stuff mactec54.

As far as not helping - since is it only mild chatter (important but mild) I am hoping the settings were close to start. Anyhow fingers crossed. Graham

[gfacer2]

Hmm now, if in autotune I set inertia ratio as presumed (calculated is the translation on the yaskawa YouTube video) then I get A. 511 error encoder output pulse overspeed error. First time I got the error on the moment of inertia test part of the tuning but on the second try I got it on the gain search portion on a relatively aggressive move.

The moment of inertia setting seems high but I have no point of reference (has varied from 385% to 415%)...

Sent from my LG-D852 using Tapatalk

[gfacer2]

OK so I reverted back to the initial (factory) settings which just had a 16 to 1 gear ratio. Ran autotune again and go the fail error. Changed the position completed width to 70 reference units from 7. Tried again and succeeded.

I've now run a comparison of the old parameters (factory) vs after autotune. Which I'll add in a second. I assume I'll add in the ones that the factory had that autotune left empty. I also assume I'll need to match up some of the controller parameters like speed loop gain etc.

Sorry - don't know why the attached PDF is previewing as a black square bit it seems fine if you open it.

[gfacer2]

Hmm, this isn't going particular well. When I combined the autotune settings and the factory settings (mainly for the metrolink settings) I get errors when homing x axis (the errors say hardware limits exceeded on y, x, z). I can massage it to finish homing but then I get a fatal following error pretty fast when I jog x.

Kind of annoying really. So what I think I will do is change the settings parameter by parameter.

Sent from my LG-D852 using Tapatalk

[mactec54]

gfacer2

There seems to be a lot of people using this control, and not all are happy customers, I think most of your settings, changes should be in the control, the Yaskawa Drives you usually set and forget, I will have to learn more about this control,

Look at the mechanical Rigidity Settings, these are the most important settings when using auto tune, they give you a good guide, for Gain settings Etc, you don't have many Parameters to play with when using Position Control, if your control is with/using step/dir

[gfacer2]

Well there are a few settings in the control, but not too many. Some I think like loop gain are in the controller but I think they need to be matched to the drive if changed as I got error galore when tried.

I've actually just made a bit of progress changing the moment of inertia ratio from factory 100% to 407% as per the autotune. It's helped a lot.

The next setting I tried is turning on friction compensation (from 0 to 85%) but that didn't work so well. No errors but motor develops a hum when stationary.



Sent from my LG-D852 using Tapatalk

[mactec54]

gfacer2

I don't think your drives have the Rigidity settings, so this won't be of any help, it seems that the Advanced auto tune supposed to do this without any other Parameters to be set, I don't have that model Drive manual so not sure how or what Parameters you have

[gfacer2]

Well - I seem to have a lot of parameters but which are useful to me I have no idea. The advanced autotune only chanced about 10 parameters. Still seem to be having a resonance issue. I thought I had it licked when I dropped the inertia setting down to 350% as it stopped dead then but I was just sitting to type this and heard it develop again.

Ugg... - on the plus side I've at least seen some improvement but on the negative its been about 5 hours of playing around with only 1 partial success (hopefully full if I can lick this resonance hum that is developing). And BTW for anyone who stumbles on to this later - I tried putting it on the tuneless feature and that shuddered the machine quite a lot. Not enough to slam on the estop but not happy.

I also tried the Y's (by manually adjusting them both) at 300% inertia (I figure I just made up) and 600 on the servo position loop gain. No real difference over just the X axis adjustment. I have found that the gain parameter in the controller is reading it from the drive on startup (syntec parameter 181/182) so that's handy to know.

[gfacer2]

Ack, just stumbled across a tuning screen in the syntec. Hit parameters and arrow right.

OK looks like I do have access to some parameters here. https://uploads.tapatalk-cdn.com/2016...c9cdae43d8.jpghttps://uploads.tapatalk-cdn.com/2016...e8fb4aaa9d.jpg

To late to play but might mean factory will have some idea of suggestions.



Sent from my LG-D852 using Tapatalk

[mactec54]

Well - I seem to have a lot of parameters but which are useful to me I have no idea. The advanced autotune only chanced about 10 parameters. Still seem to be having a resonance issue. I thought I had it licked when I dropped the inertia setting down to 350% as it stopped dead then but I was just sitting to type this and heard it develop again.

Ugg... - on the plus side I've at least seen some improvement but on the negative its been about 5 hours of playing around with only 1 partial success (hopefully full if I can lick this resonance hum that is developing). And BTW for anyone who stumbles on to this later - I tried putting it on the tuneless feature and that shuddered the machine quite a lot. Not enough to slam on the estop but not happy.

I also tried the Y's (by manually adjusting them both) at 300% inertia (I figure I just made up) and 600 on the servo position loop gain. No real difference over just the X axis adjustment. I have found that the gain parameter in the controller is reading it from the drive on startup (syntec parameter 181/182) so that's handy to know.

The main Gain is to high, lower by small amounts, and do autotune after each change, you can also do advanced auto tune, without having the control connected to the drives, this may not work though as it should as the control may have direct control over the Encoders

Yes the control may have total control of the Drives, which makes it harder to set any Parameters in the Drives

[gfacer2]

I think I'll have another try sometime when I have time.

Still, any idea what to do with the slaved axis? Driving me crazy that I can't find anything on that.

Sent from my LG-D852 using Tapatalk

[spoiledbrat]

gfacer,

old thread, but i will try anyways. i am revisiting my yaskawa sigma 5 servo tuning, for my gantry. the newer software lets you tune "with reference" (Mach3 in my case). jogging my x back and forth, i tried it with one of the x servos. the servo began to sound very good... better than ever. however, it cancelled the tuning on me because of saturation (too much torque). i suspect it was fighting the other servo. did you ever figure out how to tune?

rob