RE: [CAD_CAM_EDM_DRO] Re: Big servo motors drive recommendations

That was my thoughts Patrick about 12 months ago. I looked at the old
analogue servo cards covered in oil and crude and decided to look at digital
cards, the cheap hobby ones. I believed all the BS I was told. There are
some digital ones that are OK, ELMO do some, but they are £2500 each, Ouch.

The biggest problem I have found with all the DIY ones is it seems the
designers don’t understand how the requirements change as the motor gets
bigger to keep some viable dynamic range, and if you mantion dynamic range
they will say it’s OK the PID takes care of it. They all say it will be OK
you just have to drop the acceleration and speed a bit. They design their
cards so that the error signal gives a fixed ON time to the PWM signal.

Consider for the minute you have just started the motor and the error has
gone quite big. Reducing acceleration would reduce this. But the servo has
sent a command of lets say 50% on time to the Fets in the bridge stage. Big
motors have very low resistance, mine are about 1 ohm so we are looking at a
target current of 80 amps here. Ooops, the fets go bang or the overload
trips. So to adjust this so you can launch the motor you reduce the
acceleration and you reduce the gain so the command is less, but the drive
can only manage say 20 amps so we must cut it back so the PWM is on for
about 12.5% of the time. That’s good we’ve got the motor started, but now
very loosely we only have 12.5% of the motor’s torque available. WE can add
some I and D in the parameters to help this, but we will never get very far.
All of a sudden our massive motor has become a small motor capable of about
1/5 of it’s design torque. The analogy isn’t quite correct, but it’s pretty
realistic keeping things simple.

Now if the machine is light enough, and the motors oversized enough to allow
massive machining cuts 1/5 of the torque may be enough for hobby work. In my
case no matter what I do the Y axis won’t move fast enough to overtake a
snail.

Smaller motors have higher resistances compared with their spinning
inductance and the peak to stall torque current are proportionally much
closer, plus the lower voltage Fet’s will take a lot more peak current
abuse, so you can heavily damp the current trip to ignore that initial
inrush and all will survive. Bigger motors and heavier machines not only
have a bigger current problem but because of the weight/ mass or iron the
impulse lasts much longer enough time to let the smoke out.



Digital servo development was in it’s infancy back in the 1960’s. Back then
this problem was recognised and digital compensation methods offered to
compensate for the dynamics problem. Sadly I see no evidences that any of
our hobby guys know anything about this. In reality they haven’t thought the
control out, they have just thought PID is really simple, applied something
which worked on a bench top motor and tried to make it bigger for bigger
motors. It works OK on some big motors, not others. S curve acceleration
would dramatically help this, but we presently don’t have this luxury on
Mach.



There are some other issues with digital systems that some of the DIY cards
have addressed and some not. A digital system goes around a loop, it looks
at the error, the accumulated error, and the change compared with what it
was last time around and apply P I D multipliers to each of these. But if
you go quickly around the loop, so quick that nothing has really moved you
won’t see any change from the last time and the D part of the equation is
always useless and zero. You need to be able to adjust the loop time, or
change over a number of loops, or more complex software methods that amount
to the same. Some digital cards do this, Rutex do it, Pixie cards do it,
many do not. It is easy to see whether he card does this, if it has some
sort of loop parameter. Some PWM drives also have a motor inductance
parameter that they use to alter the PWM frequency and which could be used
to collect some partly relevant loop data.



Adjustments of the PWM frequency can help other issues. Due to the motor
inductance, when you connect power, the current flowing grows, it is not
instantenous and being able to alter the PWM frequency can help current
measurements, and limit massive peaks that occur on stalled motors.



If your analogue cards are OK and still working, I would keep them. If they
fail, find another analogue card. The people that build these bigger control
systems inherently understand the needs of bigger motors. Talk to some of
the DIY designers, talk to some of the Analogue designers, quickly you will
realise they are very different engineers especially when it comes to
experience. I will say however that the BP machines are quite light and a
known entity, there are several people claiming they are managing with
digital cards, so you may be OK.

For me I have spent enough money on useless digital cards, and countless
hours to rebuild the machine each time.



The only other good alternative is to go brusless new motors and cards.
Wayne Weedon has been looking at a lot of new brushless drives combinations
and has found some that are surprisingly cost effective. .







Dr. Mark Vaughan Ph'D. B.Eng. M0VAU

Managing Director

Vaughan Industries Ltd, reg in UK no 2561068

Water Care Technology Ltd, reg in UK no 4129351

Addr Unit3, Sydney House, Blackwater, Truro, Cornwall, TR4 8HH, UK.
Phone/Fax 44 1872 561288

RSGB DRM111(Cornwall)

_____

From: CAD_CAM_EDM_DRO@yahoogroups.com
[mailto:CAD_CAM_EDM_DRO@yahoogroups.com] On Behalf Of Patrick J
Sent: 27 May 2007 04:06
To: CAD_CAM_EDM_DRO@yahoogroups.com
Subject: Re: [CAD_CAM_EDM_DRO] Re: Big servo motors drive recommendations



I have a simular size CNC-machine. If I were to upgrade it to something
newer
than the Heidenhain TNC 155. Id like to keep the DC-servos (Max 115vDV,
10A for X, Y and 20A for Z) they have built in tachos which might not be of
any use?

Would it not be an good idea to skip analogue servo cards and go with
digital
servo cards? I dont know if digital is better but analouge sounds like
history?

Do you have any urls for those not totally overpriced analogue drives? =)





[Non-text portions of this message have been removed]