So it seems the engineer who spec'd the driver's optical isolator never
expected a user would drive a motor over some speed limit.    He saved his
company $1 per driver but not have to go  with the current generation of
Japanese parts but using instead cheap generic Chinese parts.

He made the right design and in 5 minutes of work justified two years of
his annual salary.

But if you were doing a from scratch new design and a goal was the faster
possible motor speed you not build the current machine Kit based system.
It is overly complex and has a long pipeline of to many interfaces.
If doing a from scratch design I'd cut some expensive hardware liker the
PC, the breakout cards, the Mesa card and the driver boxes.

I'd have a $3 ARM-m CPU and a $4 FPGA on the same PCB and the FPGA would
directly generate the micro-steps.   All the real time  work and the stuff
done inside the Gekko driver would've in a cheap FPGA and the ARM would
handle g-come interpretation and yes a Web Browser make a nice display
screen and runs on either a computer or a tablet or even an iPhone.

The little board with the CPU and FPGA would cost about $20 to make.  It
would connect to a "power board" that has all the high voltage stuff.   You
could either optically iso late it or not, as you are only risking a $20,

The current design is expensive and complex because it uses a collection of
parts that each try to be general purpose.  And the thing that limits
performance and reliability is all the damn wires.  If you integrate
everything from G-Code reading to micro stepping on one small PCB you
eliminate a ton of wire and cost and it's faster.

I have this same problem with my robot designs.  I can not possibly compete
with people who are willing to design PCBs.  I end up with a rats nest of
wire covering a 1/2 square foot and they build something the size of a
credit card that looks neat.

If you think about it the current machine kit based CNC controller is a
horribly inefficient design.   But it is the design that uses parts that
you can buy off the shelf.   But we pay for that convenance.   Using an
entire PC just to read g-code file and provide an operator interface is a
wast.  But on the other hand PCs can be found for free headed too the
landfill.   Then we pay Megs $100 for a $4 chip and another $100 for cables
and other "glue" and then another $100 for yet another processor in a
driver box.   All of this would easily fit on one small PCB but it would be
a very specialized PCB only good for motion control.

So, bottom line is I have now doubt at all these guys did what they said
they did.  It's just some straight forward enginerring and a willingness to
start from scratch endnote depend on off the shelf parts.



On Sat, Aug 26, 2017 at 3:58 PM, Gene Heskett <[email protected]> wrote:

> On Saturday 26 August 2017 17:16:40 Chris Albertson wrote:
>
> > I just went to Mouser.com and did a parts search.  It is really easy
> > to find low cost optical isolators that runs at 20MHz.   20Mhz seems
> > to the common rate for modern parts but if you look for faster parts
> > you can find 50Mhz without having to spend over $4 each.
> >
> > If you really need very high speed you'd use a tiny one inch long
> > fiber optic cable for isolation but that costs more.  These run at the
> > GHz range. and required specialist engineering skills to use.
> >
> > But in any case normal hobbyists can do 20MHz without need for
> > "exotic" parts.  The $2 parts from Mouser or Digikey would work.
> >
> > But can your stepper motor run that fast?  I wonder what the upper
> > limit is.  It is like a function of the motor inductance and how many
> > volts you can spell put into the motor.
> >
> I have run a bare, non-loaded 473oz nema 23 motor at a hair over 300
> kilohertz. Set for a/8 in the driver, that was a hair north of 3000
> rpms.
>
> Very very fussy about the drive and offset/duty cycle from a 2MHz
> function generator.  And the stalls are violent. Looking at the current
> flow into the drivers opto's, I was surprised I could go that fast as it
> was never truly off.  This was using about 55 volts and a DM860 driver
> set to put 4.5 amps into an 8 wire motor, wired parallel.
>
> Based on that, I came to the conclusion that the real world max revs for
> a decent low inductance stepper, weren't being limited by the rest of
> the driver, but by the speed of its own opto-isolated inputs.  Reliable
> at 750 to maybe 1500 revs. But watch for stall inducing resonances at
> fractions of that speed.  I had made some fender washers separated by
> rubber sheets whose central hole was quite tight on the drum, 9 to 15
> washers deep, and stuck them on the back shaft of the motors on my toy
> mill. Ditto the back shaft of the z motor on TLM, which trippled its
> speed. By absorbing the magnetic spring resonance, I trippled the mills
> speeds. I am tempted to see about putting something similar on the rear
> shaft of the X motor on the Sheldon as its stalls very easily at 40 ipm.
> But its crowded back there, maybe 12mm's to play with before it starts
> hitting the bed.  In the work I've done with it so far, the 36" I have
> it limited to seems more than enough to get the job done, so if its not
> broke, I don't fix it. Ditto the Z, which can move better than 100", but
> 60 seems like a good limit.  Noisy tho. Something in my shop made motor
> and thrust bearing mount seems to have a natural resonance excited by
> the steppers steps, which on a 1600 oz/in motor, are violent. I need to
> drop it to a 960. Need$ $$, works ok now.
>
> > On Sat, Aug 26, 2017 at 5:00 AM, Gene Heskett <[email protected]>
> wrote:
> > > On Saturday 26 August 2017 07:38:28 Mark wrote:
> > >
> > > At one point in the video they talk about very high frequency
> > > drivers, but you can't get much over 250 kilohertz through an
> > > opto-isolator reliably. BTDT with my own testing, So unless they are
> > > using the capacitive couplers even in their motor driver hardware,
> > > an expensive item, they are full of it.
>
>
> Cheers, Gene Heskett
> --
> "There are four boxes to be used in defense of liberty:
>  soap, ballot, jury, and ammo. Please use in that order."
> -Ed Howdershelt (Author)
> Genes Web page <http://geneslinuxbox.net:6309/gene>
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Emc-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/emc-users
>



-- 

Chris Albertson
Redondo Beach, California
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to