Dynomotion

Group: DynoMotion Message: 9139 From: ericncn Date: 2/6/2014
Subject: how to distinguish weak motors from weak frame
Hope this isn't OT.
My milling machine is clearly insufficiently rigid for the tasks I want to use it for.
I either use ridicolously slow feeds with small diameter mills or I get horrendous vibrations.

Now the question for the experts here is: how can I tell whether this is due to a soft mechanical structure, or to electronics (not properly tuned drives and/or insufficient torque motors) ?
How do you distinguish from the two situations?

I have reasons to think the motors torque is insufficient and would like to try a 1:4 or 1:5 belt drive reduction (now it's direct drive), but that conversion implies a certain amount of time and money spent and I would like to do it only if reasonably certain that's the problem to fix.

On the other side I have reasons to think it's NOT motors fault, as the power consumption of the whole machine is ridicolus, a few watts, telling me the motors are working at 5%-8% of their nominal power (or 1% to 3% of their peak power).

Ideas?  If it can help, I can post specs of the machine, motors etc.
Maybe you can tell by common sense what's wrong with it...

EC
Group: DynoMotion Message: 9150 From: Tom Kerekes Date: 2/7/2014
Subject: Re: how to distinguish weak motors from weak frame
Hi EC,

Absolutely not OT.

I think a good first test would be to capture Dest, Position, and Output of your axes while cutting.  You can then plot Position/Velocity vs following error.  This can give some information on what is going on.  What amplitude and what frequency.  And if the motor response is saturating.  If it isn't saturating then larger motors are likely to respond slower and be worse.  There is an example CaptureXYZMotionToFile.c that you can expand to include more things.  If you assign it to an m Code then it is easy to have it start capturing at a specific point in a GCode sequence.

HTH
Regards
TK


Group: DynoMotion Message: 9151 From: Tom Kerekes Date: 2/7/2014
Subject: Re: how to distinguish weak motors from weak frame
Hi EC,

Absolutely not OT.

I think a good first test would be to capture Dest, Position, and Output of your axes while cutting.  You can then plot Position/Velocity vs following error.  This can give some information on what is going on.  What amplitude and what frequency.  And if the motor response is saturating.  If it isn't saturating then larger motors are likely to respond slower and be worse.  There is an example CaptureXYZMotionToFile.c that you can expand to include more things.  If you assign it to an m Code then it is easy to have it start capturing at a specific point in a GCode sequence.

HTH
Regards
TK


Group: DynoMotion Message: 9155 From: ericncn Date: 2/7/2014
Subject: Re: how to distinguish weak motors from weak frame
Thank you!
This is great new! I didn't think that you can analyze what's happening.

But, I fear you're assuming my KFLOP can read my servo encoders.
I haven't bought the Kanalog yet, and my system has the servo loop entirely managed by the servo drives, the KFLOP only talks to the drives by step/dir.

I guess I need to have the KFLOP reading the encoders before we can go on with the analysis?

Thank you
EC

---In DynoMotion@yahoogroups.com, <tk@...> wrote:

Hi EC,

Absolutely not OT.

I think a good first test would be to capture Dest, Position, and Output of your axes while cutting.  You can then plot Position/Velocity vs following error.  This can give some information on what is going on.  What amplitude and what frequency.  And if the motor response is saturating.  If it isn't saturating then larger motors are likely to respond slower and be worse.  There is an example CaptureXYZMotionToFile.c that you can expand to include more things.  If you assign it to an m Code then it is easy to have it start capturing at a specific point in a GCode sequence.

HTH
Regards
TK


Group: DynoMotion Message: 9159 From: Tom Kerekes Date: 2/7/2014
Subject: Re: how to distinguish weak motors from weak frame
Hi EC,

Yes, that is one of the main advantages of having KFLOP close the loop.  The other option is to feed the encoder feedback back to KFLOP in parallel with the drives.  It is messy with wiring and shielding and KFLOP itself only has single ended inputs.  But for a test it can work well.

Regards
TK