Re: biqu bx 3d printer ??
On Thursday, December 23, 2021 6:41:32 PM EST David wrote:
> On Thu, 23 Dec 2021 at 00:50, Tixy <tixy@yxit.co.uk> wrote:
>
> Hi Tixy,
>
> After reading your several musings in this thread regarding
> USB verses serial interfaces for CNC machines (3D printer, etc),
> I thought I'd respond, because I think they are missing the mark.
>
> > I guess it's possible to use a USB to serial chip in the
> > printer which then talks to a serial interface on the computer driving
> > the printer. But that is very clunky ...
>
> I have a budget 3 axis CNC router and that's exactly how it operates.
> Apart from that, I'm not involved with the CNC world, but I get the
> impression that this is standard practice at all scales, for historical
> and practical reasons.
>
> I'd suggest that this is because for CNC applications, USB is not
> better than a UART serial interface, and probably worse. So CNC
> machine designers would have begun with serial interfaces before
> USB was invented, and then had no reason to change to USB.
>
> Reasons that occur to me:
>
> - USB has poorer noise immunity (don't forget that a CNC machine
> contains several motors and is often used in industrial settings
> which can be electrically noisy). And in CNC, anything going
> wrong typically produces some kind of disaster of a catastrophic,
> unsafe and expensive nature.
>
> - USB has maximum cable length specification of only a few
> metres (so wikipedia says).
>
> - USB is optimised towards higher data rates.
> In general a CNC machine receives one small file of G-code, and
> then autonomously fabricates an item. It is irrelevant if the file
> transfer takes seconds, when the mechanical job takes minutes.
>
> - CNC machines aren't mass consumer items, so they
> don't need to follow fashion.
No, time is money, so the emphasis is on what does or does not work. Sometimes
you find dependability in the strangest places.
>
This is also true, and the market is full of very expensive proprietary
solutions. Solutions that both limit what the machine can do, and that also
means that many are short term solutions when support disappears when that
vendor/maker disappears with the last $100 in the company's account. That more
often than not leaves a business that bought that $200,000 machine with 19
tons of scrap cast iron that is useless when the first $2 transistor fails.
And it may be only half paid for.
Enter, stage left, an electronics expert with an install cd for LinuxCNC in
his hand, which is free and has people all over this planet doing the
engineering needed to figure out how to rip out the proprietary stuff and
rework that dead machine, more than likely making it do things it never did
before and faster to boot. As long as the shop owner understands linux and
linuxcnc are free, but the guy who comes in to make it work and does expects
to be paid for his time and know how.
And its under very rapid development to further improve it with 10 or more
github commits a week. And we've got the cream of the coders working on it.
They have a stable version that's updated maybe annually, and the bleeding
edge, will eventually be the next version being built by a buildbot, sometimes
2 or 3 times a day. I run the bleeding edge stuff on 4 machines here and in 10
years, its bit me twice, but the fix is in the next download, or I sometimes
pull the github version and build it with my own buildbot. But I have a leg or
two up on the mechanics, as I am a CET, who worked as a broadcast engineer
since 1963, keeping a tv station on the air someplace n this great country
withthe last 18 years I worked as the CE at the local CBS affiliate. So even
at 87 yo, and an 8th grade education, the electronics involved are a piece of
cake for me.
That brings me back to "the strangest places" remark above. I wanted to see if
I could run one of my machines on an rpi3. But all it has is a plethora of
gpio pins. A swedish prof took that problem and wrote an SPI driver for us. A
famously ticklish thing but with a 3 wire com, sending data out at about 42
megabaud, the receiving the fpga output back at 25 megabaud, expands to 72 i/o
pins to run machines with. And unless the pi is also running firefox, problem
to response time is around 12 microseconds. And I was first. And it Just
Works, doing things with that lathe it couldn't do new, and 4 or 5 times
faster than I could turn the cranks.
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, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
Genes Web page <http://geneslinuxbox.net:6309/gene>
Reply to: