Machsupport Forum

Mach Discussion => Mach4 General Discussion => Topic started by: tonyfoale on November 02, 2019, 07:26:58 AM

Title: Galil vs Vital
Post by: tonyfoale on November 02, 2019, 07:26:58 AM
I have a mill with DC PM motors with tachos and linear glass slides.  The velocity loop is closed between motors and drive amplifiers which take the usual +/- 10 V velocity signals.  The position loop needs to be closed in the controller.
Currently I am running a Galil 2143 card and mach3, but I want to change to Mach4.  To take full advantage of the capabilities I need to change to a Galil 4143 or 4040 card, which is not cheap.

I have just been made aware ot the Vital Systems options, so I am looking for feedback on both options Galil or Vital?  Or suggestions as to other compatible controllers.
Title: Re: Galil vs Vital
Post by: joeaverage on November 03, 2019, 02:44:42 PM
Hi,
just as a matter of interest what would it cost to replace the existing servos with modern AC servos?

I have recently bought three Delta 750W B2 series (17 bit incremental encoder) for approx $500 each. They are superb.
I suspect they would be cheaper than a new Galill......

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 03, 2019, 03:14:37 PM
Don't those still need +/- 10v to control them?
Title: Re: Galil vs Vital
Post by: joeaverage on November 03, 2019, 07:00:04 PM
Hi,
modern AC servos are (programably) multi-mode. So, yes,  you could use analogue (+/- 10V) control them,either speed or torque modes
but also you can use position mode using the built in encoder.

I suspect if you were to retrofit that position mode is the one you would consider. Note that it renders the glass linear scales redundant.
That may be an issue to you in that linear scales 'enclose' backlash within the feedback loop whereas the rotational encoders on a AC
servo do not. If your ballscrews have no backlash then there is no real difference between the two. If there is backlash then the scales
enclosed in the loop are probably a feature that you would not want to miss on.

Note that you could use the linear scales with AC servos but it would still require an analogue capable controller and any savings you
hoped to make vanish. For instance even, the cheaper of your two options, the Hicon with the required analogue activations is $1200 whereas
a step/direction motion controller like an Ethernet SmoothStepper is $180. Thus the cost saving would pay for two Delta 750W B2 series servos AND drives.
It makes no sense to buy new servos only to then have to use an expensive controller.

The current trend in servos, and has been for twenty plus years, is ever smarter servo drives. In  the early days servo amps were just that 'amplifiers'.
In the very earliest of days there was no tacho feedback.....then came tacho feedback, the came step/direction control with the servo mounted encoders
being monitored by the drive and closing the loop. Gecko 320's are an example of these, brushed DC motors but with the feedback loop completed
by the drive NOT the motion controller. Nowadays with field oriented control has given rise to brushless servos.

More recently still with Ethercat, Profibus, CANOpen and the likes the servo drive assumes responsibility for the motion control of its own servo, this
is called 'distributed motion control'. Its coming whether you like it or not. I say 'go mad now and beat the rush'.

You may have noted on the NFS Forum video banner a Matsuura VMC machining a part. What may not be evident is that its Ethercat controlled. Interval Zero
have contributed a real-time core and plugin to a basic Windows PC and Kingstar have written an Ethercat module for Mach4.

I'm not suggesting that you leap quite that far yet, Ethercat is at the leading edge of development for Mach4 even if it is well established in industrial applications.
None the less, I am suggesting that you move to a natively step/direction controller and AC servos. If you absolutely require linear scales to be included in the feedback
loop then its not an economic proposition.

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 04, 2019, 04:21:42 AM
Craig,

Firstly, let me thank you for the time that you have spent on this.

If I need to feed these new controllers with +/- 10 V surely I would still need something like the Galil card to supply that.  The +/- 10 V is a velocity signal. So if the position loop is closed in the controller how does it know where to stop, if it is not sent a position signal.

I must admit that I had a bit of tunnel vision and did not really consider changing the motors.  I have had the mill for 10/12 years which had been owned by a friend since new.  It came fitted with an Anilam Crusader CNC from new.  It came to me with a complete spare set of motors, drivers and slides.  However, the controller was way outdated and not user friendly.  So my initial motivation was to replace the controller and that is what I focused on, changing the "good" stuff was not considered. 

At the time there did not seem to be many options open to me and I settled on the Galil and Mach3.  The Mach4/Galil plugin was not available then.  I never liked Mach3 and recently it gave me a lot of hassle so I am revisiting the idea of using mach4 with the recommended higher spec card.  Again I only considered the controller side.  However, your comments have widened my horizon somewhat.

If I replace the motors and drivers, as you suggest, then is a PC based system the best way to control them?  Would I not be better off getting an all in one controller/user-interface panel?  To stay with a PC system I would want to make a switch panel, including an MPG and  speed and feed controls to get away from constant mouse usage, then there are mach licences etc.  I would need to replace my 32 bit computer with a 64 bit machine to run the Galil libraries etc.  Without those cash and time expenses it would not be too onerous to go for a complete controller panel.

For a giggle I attach a pic of the old Anilam controller.
Title: Re: Galil vs Vital
Post by: joeaverage on November 04, 2019, 07:10:18 AM
Hi,
there are quite a few questions in your post.

Quote
If I need to feed these new controllers with +/- 10 V surely I would still need something like the Galil card to supply that.

You are correct you would need a Gallil or something like it IF you persist in feeding the servo amps with an analogue signal.
The AC servos I'm recommending are step/direction, not analogue at all. You can thereby avoid the need for a feedback capable
controller.

Quote
So if the position loop is closed in the controller how does it know where to stop, if it is not sent a position signal.

That's what the linear scales are about. They not only drive the DROs but also supply the controller with the current actual
machine position and the controller will compare that to the commanded position and cause the motors to drive in a direction
and speed to match the commanded position exactly.

The AC servos I'm recommending don't need  linear scales because it has a built in encoder. Thus if the position is commanded
to move 10mm to the right and the machine has a 5mm pitch ballscrew the ballscrew must rotate two times. If the encoder
has a 10,000 count per rev encoder (entry level these days) then the trajectory planner/motion controller will issue 20,000
pluses and the servo drive will follow that command and monitor the encoder so that the commanded position is achieved
exactly. In this instance its the servo drive which 'closes the loop' rather than the controller.

Modern servo drives do a very VERY VERY good job of closing the loop, easily as good and in many cases better than
Gallil simply because the manufacturer is making a drive for his servo and he knows them both extremely well. Galil on the other
hand have to accommodate such a wide variety of motors and can't be great at all of them.

Quote
If I replace the motors and drivers, as you suggest, then is a PC based system the best way to control them?

This is a bit of a loaded question. I think it has less to do with whether you replace the servos or not but rather is  'PC the right
way to control a CNC machine?'. The shortest answer is NO. The best way is with a dedicated hardware controller like
a late model Fanuc or Seimens 840. These controllers are worth $20,000 plus. The question is whether you have the budget
to buy a dedicated controller or you can afford a much more humble PC.

A suitable but low powered PC for Mach4, an Ethernet SmoothStepper and a MB03 breakout board, and a Mac4Hobby
license would  cost under $2000 new. That leaves quite a bit of budget for controller panels and pendants etc.
I run my mill on a Mini-ITX dual core Atom board without a graphics card, as the saying goes, 'it couldn't pull the
skin off a rice pudding' but it runs Mach fine. Its 32 bit by the way, Mach4 runs on either 32 or 64 bit without demur.

If you have budget for a dedicated controller like Fanuc or Siemens and a motion controller like Gallil then why mess about
Mach4 or in fact any PC based solution? That you are on the forum suggests that you don't have that sort of budget.
Mach4 is one of the Windows PC based software solutions that can achieve good, surprisingly good, performance relative
to dedicated hardware controllers worth ten and more times as much.

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 04, 2019, 08:34:37 AM
Craig,

Thanks again.  As I said you have widened my horizons towards other and probably better solutions which I will consider without rush.

You are right, it is doubtful that I would be on this forum if I had a high end budget.

The reason that I would need a 64 bit computer is to use the Galil library files if I stayed with the original plan.  They are supposed to work on 32 bit machines but do not. I know Mach4 itself will run on 32 bit and even with Win XP.  When I asked about dedicated control panels I was thinking about the lower end offerings at around $1000 which o/p step and direction data.  Would those signals feed directly into the new servo drives or is an Ethernet SmoothStepper and a MB03 breakout board still required?
Title: Re: Galil vs Vital
Post by: joeaverage on November 04, 2019, 10:16:10 AM
Hi,

Quote
When I asked about dedicated control panels I was thinking about the lower end offerings at around $1000 which o/p step and direction data.  Would those signals feed directly into the new servo drives or is an Ethernet SmoothStepper and a MB03 breakout board still required?

The only dedicated controller I have seen in that price range was a POS Chinese controller fitted to a customers plasma table.
Between the controller and the stand-alone THC unit they were the Achilles heel of the whole setup. I convinced them to ditch
the Chinese crap and get Mach4 and a Hicon, as at that time the Hicon only had THC among the Mach4 motion controllers.

It must be said that the Chinese controller produced step/direction pulses directly to the Mitsibishi servo drives fitted to the
table.

The ESS (or the Hicon come to that) are specifically for Mach (3 or 4) loaded PCs. Ergo they are not required nor would work
with ANY other control software. The MB03 is a breakout board designed to work specifically with an ESS and will only
EVER be partnered with an ESS.

The cheapest quality controller I've seen advertised is a Siemens 803 (three axis) at around $3500 USD (new).

Don't be duped by the Chinese rip-off Siemens look-a-likes......you MUST establish it is the genuine thing BEFORE
parting with a cent.

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 04, 2019, 10:45:24 AM
Craig,

Many thanks, I shall heed your advice.
Your comments indicate that you do this sort of thing for a living or that you have experience with many machines.
Title: Re: Galil vs Vital
Post by: joeaverage on November 04, 2019, 10:57:56 AM
Hi,
I did at one time deal with a few plasma tables but not CNC machining centers, so excepting some experience gained
from customers plasma tables my knowledge is as a hobbyist only.

Craig.
Title: Re: Galil vs Vital
Post by: mcardoso on November 04, 2019, 11:11:29 AM
My machine is running on AC servos using Step/Dir (actually quadrature) signals from an Ethernet SmoothStepper (ESS) from Warp9TD and Mach4. I am extremely happy with the system. I think you will find AC servos to offer higher performance compared to DC systems.

I think JoeAverage has done an excellent job describing the system, so I won't get into that, but one thing to consider is selecting a servo drive which has dual loop control using load feedback. In this case, both the motor encoder AND the glass scales would be wired into the servo drive which handles the position, velocity, and current loops entirely on its own. You only need to give it a command (Step/Dir for example) and it handles everything else.

If you use an ESS, you can easily bring the encoder outputs from the drive back to Mach 4 for display purposes. As far as Mach 4 is concerned, it is driving an open loop stepper motor, but you really have full closed loop control at the drive.

Not every servo drive can do this, but I know the new Kinetix 5100 from AB has this feature as well as select drives from Delta, Omron, Baldor, and likely most big name suppliers. Give them a call and ask what their offering is for standalone drives with dual loop load feedback.

https://ab.rockwellautomation.com/Motion-Control/Servo-Drives/Kinetix-Indexing-and-Stand-Alone/Kinetix-5100

https://www.deltaww.com/Products/CategoryListT1.aspx?CID=060201&PID=23&hl=en-US&Name=ASDA-A2+Series


If I had a nicer machine, this would be the route I would take!
Title: Re: Galil vs Vital
Post by: joeaverage on November 04, 2019, 03:20:33 PM
Hi,
yes the Delta A2 series have dual reading encoder inputs, you can have the best of both worlds, a 1,280,000 count per rev rotary encoder AND
your existing linear scales. The A2 series commands a premium of several hundred dollars over the 'standard' B2 series.

As I have posted previously, modern AC servos and their drives just leave all the old DC servos in the dust!!!

Craig

Title: Re: Galil vs Vital
Post by: joeaverage on November 06, 2019, 12:01:19 AM
Hi,
just saw this:

https://www.ebay.com/itm/ECMA-C10807RS-ASD-A2-0721-L-DELTA-A2-750w-3000rpm-AC-servo-motor-driver-kit/172283162086?hash=item281ce06de6:g:TkYAAOSwIsFcpYRy (https://www.ebay.com/itm/ECMA-C10807RS-ASD-A2-0721-L-DELTA-A2-750w-3000rpm-AC-servo-motor-driver-kit/172283162086?hash=item281ce06de6:g:TkYAAOSwIsFcpYRy)

It would appear that the A2 series is NOT that much dearer than the B2 series. I wish I had known that a month or so
ago!

Craig
Title: Re: Galil vs Vital
Post by: mcardoso on November 06, 2019, 11:11:25 AM
Those are some solid prices for a really good drive!  I see the drives selling used for $150-250 on ebay.
Title: Re: Galil vs Vital
Post by: joeaverage on November 06, 2019, 01:07:55 PM
Hi Tony,
with this new capability this would be my recommendation for your original question:

Replace your existing DC servos with 750W Delta A2 series servos and drives. They would operate dual loop, that is
with their own built-in encoder AND your existing linear scales.

The cost would be:
3 x Delta 750W A2 servo/servo drive            $1524.75
ESS                                                             $180.00
MB03 Breakout Board                                   $180.00

Total                                                           $1884.75  Plus shipping and taxes if any.

That solution would be cheaper than a new Galill but rather more than a Hicon (with analogue activations) but have the advantage
of crispy new servos.......I can promise you they will blow your socks off!

Craig
Title: Re: Galil vs Vital
Post by: joeaverage on November 06, 2019, 01:15:02 PM
Hi Tony,
should have added the price of a drive programming cable. I bought a USB to IEEE1390 for $62 including shipping.
Alternatively, and a bit cheaper, RS323 to IEEE1390 are about $35.

I regard a programming cable as all but essential if you are going to use late spec Delta servos.

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 06, 2019, 01:59:06 PM
Craig,

I revisited your responses from the beginning and I note that you mentioned the Gecko 320X.  I just had a look at those and they appear to be suitable.  OK maybe not state of the art but my needs are not production nor customer based.  Purely for making some parts for my own classic racing motorcycles.  I guess that the Gecko performance is short of what the Deltas would produce but otherwise would they do what I need if I kept the original motors?  I guess that I would still need the ESS and breakout boards?
I had previously associated Geckos with stepper motors and so I never considered them earlier.
With any of these options do I feed other inputs like an MPG in through the breakout boards or.....?
Title: Re: Galil vs Vital
Post by: joeaverage on November 06, 2019, 02:42:29 PM
Hi Tony,

Quote
I revisited your responses from the beginning and I note that you mentioned the Gecko 320X.  I just had a look at those and they appear to be suitable.

Sorry to burst your bubble but I don't think they are suitable. The Geckos  G320 are expecting a rotary encoder that must be directly and positively linked to the servo motor
shaft. What you have is a linear scale that is not absolutely and directly linked. I suspect that that flexure/backlash/lost motion between the axis (as measured by the
linear scale) and the servo motor shaft would render the G320 feedback loop unstable.


Craig
Title: Re: Galil vs Vital
Post by: mcardoso on November 07, 2019, 10:56:24 AM
Or install an encoder on the motor shaft if it doesn't already have one. AMCI and US Digital offer low cost low profile encoder options. The glass scales wouldn't be used in this case.
Title: Re: Galil vs Vital
Post by: tonyfoale on November 07, 2019, 12:07:43 PM
Yes I was thinking about that also.  If the tachos were not used then there would be plenty of room.
The motors and ballscrews are connected by a toothed belt, would it be better to put encoders on the motors or ball screws?  On the motors would be the easiest.
Title: Re: Galil vs Vital
Post by: joeaverage on November 07, 2019, 12:31:54 PM
Hi,
if you are going to fit encoders then fitting direct to the servo shaft is preferred as it will result in the most stable
position control loop.

What you are proposing is to fit rotary encoders so the servos may be driven by G320's which are step/direction input
drives, which are in turn commanded by an ESS/BoB as a step/direction controller ALL to avoid buying an
analogue controller?

I suspect that you are adding a great deal of complexity in the conversion for little gain. The money saved after you buy three
encoders, three G320's, an ESS and a BoB vs buying one Gallil or Hicon is likely to be small. Does that small gain cover the
added work and 'risk' of the conversion on the lines you have proposed....I'm dubious.

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 07, 2019, 05:50:20 PM
I was just considering options and learning.  That was not a definite proposal that I had real intentions of following.
Title: Re: Galil vs Vital
Post by: joeaverage on November 07, 2019, 05:59:22 PM
Hi,
its worth considering but I would have thought that the risk of buying all that gear only then to find out it doesn't work
or work well is a bit much. Hey, if you could buy just one drive and an encoder and experiment before spending big bucks....maybe.

Maybe someone else out there has tried such a thing. Their experience would be invaluable.

To my knowledge no-one has been able to get linear scales to work WITHOUT using an analogue controller, I could be wrong but thats
my understanding.

Craig
Title: Re: Galil vs Vital
Post by: tonyfoale on November 07, 2019, 07:08:25 PM
Craig,

I thought that the main advantage of the A2 over the B2 was that the A2 could process the motor encoder and the linear encoder whereas the B2 could work with the motor encoder only.  Or did I misinterpret your earlier comments?
Title: Re: Galil vs Vital
Post by: mcardoso on November 07, 2019, 08:33:50 PM
That is correct. It is a feature that very few use (outside of professional machine tools) but is very powerful.

This usually closes the velocity loop using the motor encoder and the position loop using the glass scales. This way you get high dynamic response with tight position control.
Title: Re: Galil vs Vital
Post by: joeaverage on November 08, 2019, 03:19:26 AM
Hi,
one of the principles on which field oriented control rely is accurate information of the angular position of the rotor,
ergo the use of an encoder or resolver. Any flexure or backlash between the rotor and the encoder will seriously
degrade the maintenance of quadrature flux angle thereby degrading motor performance.

Field oriented control on which AC servos rely is an interesting topic, I suggest a Google search 'field oriented control by
Texas Instruments'. It will reveal a couple of excellent introductory videos on the subject. The upshot is that ALL AC servos
capable of accurate position control must have an integrally linked encoder/resolver. That does not preclude a secondary means
of detecting position but the servo drive MUST have direct access to rotor angle with zero lash/lost motion. The encoder
in the B2 series doubles as position sensor (rotational position is converted into linear position through a low lash/zero lash
ballscrew) and rotor angle for the field control loop to operate.

The A2 series is dual reading, thus it uses the encoder for field control but can 'load sense' using the auxillary encoder/linear
scale channel. Very crafty indeed. Had I realized that for as little as $50 extra each, I would have bought them instead of the B2
series examples I did get. Having said that I have no need of the 'extra fruity bits', I doubt I'll exhaust the B2 series capabilities
let alone the A2 series!!

Craig