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

Hi Zafar

That sounds good sense especially when you have the Pixies.

My R2020’s worked brilliantly on my X axis. Got the encoder noise issue we
all have problems with, and was running 400 to 500mm/sec/sec acceleration
and 8m/min rapids. With Quantum I could do 700 to 800mm/sec/sec and 16m/sec.
What was nice compared with others was the precision, most of the time the
error was less than 30 counts to an impulse response, other cards I have
tried the error is thousands to a slow acceleration, they would not work for
an impulse response. The PID in Rutex is good, and they have some form of
dynamic current compensation, but it is not quite enough.

On my Y axis most of the time the Rutex were OK for 400mm/sec/sec and 8m/min
which was workable, but every now and a gain Y would go real grunchy, I
initially thought my Y screw had collapsed so I rebuilt it, rebuilt the
motor brush kit. Then looking on a scope I realised the Rutex was going
overcurrent, it would foldback, and part of the bridge would not properly go
into conduction. We were also getting massive errors during machining of
light plastics, a hole would be smaller in the Y, but when you stopped the
machine to measure the position, the I parameter would cause it to follow up
to zero error. My machine was originally designed to machine cast at hi
speeds for it’s time and with virtually unstallable motors, I don’t need
that amount of power, but I do need more than DIY digital cards have
offered.

You know at present based on what I have learned over the passed months in
this area, I do really wonder whether I should be designing my own digital
card. This is not my subject field it is one of my brothers speciality areas
and one we have a lot of books on, but it hasn’t taken long to pull all the
faults out of what we have, and find effective theoretical solutions for
them. Presently I have so much other design work on the table it will be a
long time before I could play with this though, and my order book seems to
continually fill quicker that I can get the stuff out the door, so don’t
expect anything in the near future.



Regs Mark



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 Zafar Salam
Sent: 27 May 2007 07:31
To: CAD_CAM_EDM_DRO@yahoogroups.com
Subject: [CAD_CAM_EDM_DRO] Re: Big servo motors drive recommendations



Mark,

Thanks a lot for the detailed reply. I am already using Rutex 2020
drives on smaller machines after putting encoder cables in armoured
sleaves and putting noise filters on motor outputs. These are giving
me excellent performance. I am getting 5 m/min rapids on Bridgeport
Series I retrofits with their original SEM motors and new
differential encoders. I haven't tried it yet but I guess with
scurving in Quantum I can get higher than 200 mm/s/s acceleration
which I am getting with Mach3.

For the bigger machine, I was also thinking about using commercial
analog amplifiers with Pixie drives. I happen to have three Pixie
drives lying around for quite some time. It's time to put those to
work. Could you post links for the cheaper analog amplifiers you
mentioned in another message.

Zafar

--- In CAD_CAM_EDM_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>
DRO@yahoogroups.com, "Mark Vaughan" <mark@...>
wrote:

>
> I have a mill with motors not quite so big. Typically mine would

be 140 to

> 150V, 8.5Amp continuous unless I fan cool them in which case the

continuous

> is about double, and 46 amp peak.
>
> They are fitted to quite a weighty machine so acceleration

currents can be

> quite hi for quite a long length of time.
>
>
>
> I started with Rutex R2020 drives. These have a good PID control

that uses

> torque / current in the equation rather than a PWM figure.

Consequently the

> dynamic range is fairly good. However they suffer from noise

issues, and

> slow current trips (fet's pop before the trip). They have

presently been

> withdrawn from sale for redesign. At one time I thought they were

good, then

> after problems I changed my mind, but in comparison to others they

have got

> more right I think than a lot of the competition, and I was

machining with

> them for several months until one developed senile demeter and

lied about

> it's position. With no replacement available I tried another.
>
> CNCTeknix Tek20's, lot's of faults here, insufficient voltage

rating on some

> of the fets, bad current trip design, PID is a PWM figure with no

current or

> speed compensation so dynamic range is not viable for a big motor.

I.e. You

> have to keep the gain so low to avoid current tripping when the

motor is

> slow that you get significant follow errors when you are into

machining

> speeds let alone rapids. Peter at CNC teknix has done everything

he can to

> help, but a simple fix isn't going to do it. I was also assured

Tek10's

> would work at reduced voltage so I tried them as well but these

also suffer

> from the same problems. I am sure they work OK on small motors but

forget

> about it at present for your size of motors.
>
> I looked at several other DIY hobby style cards and have come to

various

> conclusions but generally they are not suitable for big motors, or

do not

> instill enough confidence in me to want to waste another £500+ on

useless

> cards..
>
>
>
> I have now have spent several months researching servo drives for

big

> motors. Virtually all the hobby drives work their PID control to

give a set

> PWM. The trouble is at low speed with a big motor you only want a

little

> amount of PWM to avoid hitting the peak current, but as the motor

speeds up

> you need more gain and on a big motor PID into PWM just cannot

achieve that.

> You have to have very low acceleration and gain to get it to start

moving,

> and then you can't get any machining speed, or you have to delay

any current

> trip and you risk going above peak currents which could damage the

motor.

>
> There are two solutions with a basic digital control you can use

the encoder

> signal to give details of motor speed and apply this to the PID to
> compensate for speed. This is a digital technique dating back to

the 1960's

> and works well, either by effecting the pwm, or by producing a

constant on

> time and phase locking the pwm frequency to the encoder signal.

But no one

> in the DIY sector seems aware of the technique. The second

technique is to

> consider the PID output a torque command as it should be and use

that to

> regulate current to the motor, Rutex attempt this in their R2020,

but most

> DIY Step direction controllers do not properly seem to do this.
>
> If however you look at conventional analogue servo drives (+/-10V,

or +/-5V

> ) as used in commercial machines these take the voltage input

command and

> use it to directly vary motor torque or current. Most even use the

tach

> signal to compensate for speed. They can be much more expensive

but when you

> see how much control is on a card you know why. In the UK I have

been

> offered analogue cards for my machine for £250 to £350 each. You

may even

> have a set that came with the machine and still work, these

analogue cards

> seem far more bullet proof than the DIY ones.
>
> Then we need to be able to control these cards with a

step/direction signal.

> There are two firms that I know of that produce conversion cards

for these.

>
> Rutex, though I think their card may be about to go out of

production. It

> will take differential encoder signals, but could have some of

Rutex's other

> noise issues though early cards are said to be OK. If Rutex's PID

is like

> the R2020, it takes care of the typical problems of digital PID

control,

> with variable loop times to allow for differential sensitivity.
>
> Skyco Pixie100 cards. These only have single ended encoder inputs,

for me I

> will use differential encoders with buffers since I need to take

encoder

> signal on to other machine controls. The PID looks good and has a

variable

> to be able to set it so the differential is only read after a

number of

> times around the loop, again compensating for normal digital PID

control

> problems where the differential can be missed in a high speed

loop. With

> Pixie cards you also have two sets of PID parameters so you can

set one set

> when the machine is idle to keep the motors cool, and a more

aggressive set

> for machining. They look very good and so far all reports have been
> excellent and are $60 each.
>
>
>
> I have looked at some step direction current mode controllers that

were

> impressive, but at £2500 each my machine won't see any.
>
>
>
> I hope that all makes sense and saves you the cost headache and

lost hours I

> have experienced playing with hobby DIY cards.
>
>
>
> Regs Mark
>
>
>
> 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_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>

DRO@yahoogroups.com

> [mailto:CAD_CAM_EDM_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>

DRO@yahoogroups.com] On Behalf Of Zafar Salam

> Sent: 26 May 2007 16:11
> To: CAD_CAM_EDM_ <mailto:CAD_CAM_EDM_DRO%40yahoogroups.com>

DRO@yahoogroups.com

> Subject: [CAD_CAM_EDM_DRO] Big servo motors drive recommendations
>
>
>
> Hi,
>
> A mill I am going to retrofit has rather big dc brush motors.

These motors

> are 150V, 12A cont. and 125A peak current. Which servo amplifiers

do you

> guys recommend for an application like this.
>
> Zafar

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