[matttargett4]

Hi,

I have an old esab gas cutting gantry which i have modified for cnc, i doubled up the motors on the x axis and left the y alone.

I got asked to cut some really fine parts the other day and found my quality on really small parts at high speeds was poor so i satrted studying my motion more, normally i cut large parts at low speed so i never see this.

The motors are electrocraft e586's, they appear to be quite high inertia, initially i found that if i came under 1500mm/min then my motion results improved alot, faster than that and the lines were a bit like saw teeth, if i reduce my acceloration vales from 0.040g down to 0.020g i saw the quality improve at higher speed, mainly because even though the feedrate was set higher it never acheived it.

I then started looking for backlash, on my x axis i found i had to command a move of 0.05mm to see a response, on the y it was more like 0.03mm, is that alot? or nothing to worry about?

It seems as though just jogging the machine around at high speed if if is allowed to reach full speed then it will stop smoothly but if full speed is not reached before it is stopped then the machine oscillates to a stop, is the backlash causing this?

I also read some stuff about inertia matching so i installed the motioneering program and mdelled my motors and mechanism as well as i could and it appears as though the motor and gear ratio are almost spot on, if i increase or decrease the gear ratio the torque and power requiremnets seem to go up, presumably this is the sweet spot?

Where do my drives fit into this? I have whale drives from cncdrives and ive never liked the tuning software, i find the usb connection troublesome and have never seemed to really get the system right, do the drives make a difference to the system?

I hope you can give me some clues

Matt

[Al_The_Man]

I retro fitted and old Esab gas/trace cutting http://www.cnczone.com/forums/cnc_pl...le-2.html?pp=8
I used some high ratio gear boxes due to the size/weight of the gantry.
Al.

[matttargett4]

Hi al,

Thanks for the reply,i d be really interested to hear more details of what motors and gearboxes you used, what drives and also what results you managed to achieve, id be very greatful if you could help me improve my machines performance

I also noticed you have you encoders mounted so they drive from the rack no mounted to the motor, presumably this improves accuracy? How does this fair if backlash is introduced?

Do you think my 0.05mm backlash is enough to worry about? it appears to be in the gearbox/clutch mechanism from the esab machine, does it justify tearing apart and improving or not?

Matt

Matt

[Al_The_Man]

I used a Acroloop motion card with Acrocut s/w.
A-M-C analogue drives in torque mode.
The motors were BLDC 6000rpm so I used 100:1 Bayside gear box which gave around 500"/min.
The encoders were on the motors, you may have confused the air cylinders used to lift the pinion off of the rack so the operator could manually re-position the table, the end user wanted to retain this feature due to the 100' length of the table.
Al.

[matttargett4]

Thanks al

What kind of power/torque were the motors rated at?

Matt

[Al_The_Man]

IIRC they were some Aerotech motors we had on hand, BM250 5N.
Al.

[matttargett4]

Hi al

Do you have any suggestions as to how i can work out what is limiting my performance, i have been testing with a pen and paper but i have no idea where to go next

Matt

[Al_The_Man]

What kind of controller are you using?
I have used motion cards in the one's I have done, such as Galil motion and Acroloop, the drives with these I use in the Torque mode with no feedback to the drive, just the encoder to the controller to close the PID loop.
.05mm does not seem excessive for a gantry?
Al.

[matttargett4]

Im using whale drives from cncdrives.com, i use mach 3 through a pp on an old pc, my error analysis software for the drrives shows an error of only a few steps, i foget my resolution but its less than the0.05 backlash.

One thing i have noticed when tuning the drives is i seem to struggle to make the system over damped, i can make it critical with the max setting for damping and quite a small gain, an increase in gain makes the system unstable. Does this give any clues?

Matt

[gridley51]

you may have confused the air cylinders used to lift the pinion off of the rack so the operator could manually re-position the table, the end user wanted to retain this feature due to the 100' length of the table.
Al.

Hi Al
did the owner retain this so the operator could manually push the gantry to his start point on the plate and when re-engaging the drives use that position for the reference point?

[Al_The_Man]

Yes,they were into bridge structure building and wanted to do it this way for some reason.
Al.

[matttargett4]

Al,

In your opinion does the motion controller make an appreciable difference to performance over the system i am using?

[Al_The_Man]

In your case, the drive is the motion controller, this is where the PID loop is closed.
I can't really give a definite opinion as I have no experience with those drives and do not usually use step/dir control, the one I use are ±10vdc analogue torque mode drives.
But I would have thought you should get decent results with them?
Al.

[erikjgreen]

Im using whale drives from cncdrives.com, i use mach 3 through a pp on an old pc, my error analysis software for the drrives shows an error of only a few steps, i foget my resolution but its less than the0.05 backlash.

One thing i have noticed when tuning the drives is i seem to struggle to make the system over damped, i can make it critical with the max setting for damping and quite a small gain, an increase in gain makes the system unstable. Does this give any clues?

Matt

I've seen essentially the same behavior in the cncdrives DG2S-16035 servo drives.

So far I've tuned to what (like you) I think is a critically damped level with a small gain, about 15% of max d, and i set to 0. If I increase gain much the system is unstable, and increasing d doesn't buy me anything, and after a certain point starts the system oscillating (near 90% of max).

I did find an old tuning guide by the maker of the drives that may give me some insight. The problem with the old guide and the much better written but missing some information new guide is that english is a second language for him, and although he probably speaks it better than I do, some technical concepts don't seem to come across properly, so I may be missing something about tuning these drives.

Erik

[matttargett4]

Interesting

Sort of glad to hear someone else has the same trouble,its likethe damping is insufficient and the reason the following error is large is because the gainis not sufficient to keep it low, als document appears to confirm this theory suggesting acel value will increase following error if gainis insufficient, maybe this means a motion controller would improve the situation, what others are available which would work in the same way? Maybe a smoothstepper?

[erikjgreen]

In my setup I have the drives connected to a Mesa I/O card, which along with linuxcnc does high speed pulse generation like a hardware motion controller would.

When tuning the drives, I use ServoConfig software from cncdrives, and I'm assuming you do as well. All the "pulse" generation for testing is internal to the drive, so if you're seeing the same problems I am tuning, it's an issue with the PID setup inside the drive.

The only interface to a motion controller is the step/direction interface. I don't think the drive cares what generates the step/direction pulses... that's its main advantage. Most of the Cncdrives products are made for simple interfacing to parallel ports or pulse gens, so they can be used with simple control systems.

There are advantages and disadvantages to this, of course. Personally after dealing with the DG2S drives, I would use them again for certain projects, those where I didn't care to close the PID loop in the controller or where I wanted a simple control setup, or if I was converting from an existing stepper control system. Plus the fact that they can run these huge old DC servos is a big plus.

Erik

[matttargett4]

Hi erik

Yes im using servo config to

What i was meaning was what other motion controllers would close the pid loop outside the drive and allow for perhaps better tuning flexibility, although probably the question is acedemic since im in no position to throw more monet at the problem.

I am also running a gcode program and watching the following error on the errror viewer window, trouble is on a small part its all happening so fast i cant hope to see what the max value is, a peak hold function would be great, any ideas?

I emailed cncdrives yesterday asking them to take a look at this thread, evidently no luck yet

Matt

[erikjgreen]

I'm waiting on a reply from Balazs myself at the moment... I'm guessing he's either busy with his new drive release (DG3S) or his internet is down again (they do that to him a lot in Hungary).

If you look at the diagnostics tab in servoconfig, it does have a number showing max error since power-on or reset (can't remember which ATM).

Erik

[matttargett4]

Thanks alot erik

I will try retuning with als notes and check the max error