headers
gene heskett | 10 Feb 15:51

One LinuxCNC lament

Greetings guys;

Sort of early for me, retired old fart etc, doesn't live by a clock, but I 
did make a good board last night!

I had pcb-gcode drilling holes only halfway through the board, so they 
would meet in the middle and ANAICT the error is not over a thou or so.  
How sweet it is!

But one thing, while working on this, that bothered me quite a bit and 
slowed this registration fine tuning down, was that a function file like 
the one I use to probe for tool lengths at each tool change, is not shown 
in the gcode window as it runs, which made me resort to (msg, etc) and 
(debug, etc) statements in the "o<subname> call" files to see what it was 
doing.

Was this omission in how the gcode in a called file is displayed a 
conscious choice, or is it something that could be called a bug and could 
be easily fixed?

I was running the next to latest master-rt branch last night.  I see it was 
updated and I did update LinuxCNC after I was finished last night.

I will see if I can get pix of the board posted on my page in the next day.  
While I am editing that, should I change the emc references to LinuxCNC 
just to keep the other guys off my back?  Seems like it should be done for 
consistency.

Cheers, Gene
--

--
(Continue reading)

Permalink | Reply | 0 comments |
headers
gene heskett | 9 Feb 00:30

Another silly Q? And needle threader based probe report.

Greetings;

Do I need to have run the machine to X,Y=0.000 before I program a G10 L2 P2 
X#<offset>  to set G55 to run at -0.1000 from the G54 positions?  Or can I 
do that anytime the G5x mode is not in that specific mode?

Needle threader probe:

Is working quite well, I even went Z-0.050 from contact, disconnected the 
probe and drug it across the pcb while it was turning 450 revs without 
leaving more than a slightly polished track in the pcb copper that I had to 
look really hard for. Another board wrecking problem solved.  One could say 
that plan B is a huge success compared to plan A.  ;-)

Its a bit critical to bend straight enough when the touch hole is only 
1/16th inch, a 3/32 hole would be just as functional and less critical to 
center up over before I run the holefinder routine.  In that event, 
drilling .1" into the pallet might be advantageous in case I was really 
good at centering over the hole first so the probe didn't bottom in the 
hole as that would make it wobble worse and fatten the effective diameter 
of the loops conical contact form when its spinning.  OTOH, if it does, at 
least it won't have to run very far to find the 4 corners of a supposedly 
round hole.  ;-)

I haven't tried the capacitor across the contact idea to stretch the closed 
detection, but I believe I probably should, enough to make any contact that 
discharged the capacitor look like at least a 10ms contact.  I have had an 
occasion or 2 where the noise as I tried to move away enough to clear has 
caused a probe event while jogging message, so that part is a bit critical, 
but I am not damaging the board, so I can now make lots more mistakes
(Continue reading)

Permalink | Reply | 0 comments |
headers
gene heskett | 8 Feb 21:06

Can I bounce this idea off the list?

Greetings all;

I have now managed to wreck 3 or 4 half completed attempts at this board by 
typu's in my code driving the tapered hole probe too far, so I have turned 
to the contact probes construction.

1. Assume the spindle carrying the ground is turning 2-400 rpms so that any 
runnout is canceled as the probes end shape is then a cone shaped envelope 
described by the probe as it turns, and its turning fast enough that the 
probing motion has only advanced by maybe 0.0001" or less in the time it 
takes to complete a revolution.

2. The response to a true from the probe is at worst the servo threads lag 
to get the stopping initiated.  So I do 2 probes, one at say 4ipm, back up 
a bit, 0.010 maybe, and redo it at say .2 ipm for a final G38.2 snapshot 
capture in $5361-5369.

3. The current involved is a TTL pulldown, maybe 3 mills max, so electrical 
erosion of the contact isn't a short term problem.  This is I suspect true 
even if a add a small capacitor across the contact to stretch a momentary 
contact out long enough for reliable first contact detection.

4. What would be wrong with taking the approach of making the probe contact 
so flimsy it cannot damage the board even if driven sideways out of the 
hole in the pcb and across the copper while it is still spinning?

Rummaging around in the missus's sewing stuff last night, looking for a 
suitable flimsy straight pin to cabbage, I came across a couple of her 
needle threaders, those ultra teeny little springs bent in a V that you 
stick thru the eye of a needle, run the thread through the loop of the V
(Continue reading)

Permalink | Reply | 16 comments |
headers
Oscar Chaides | 8 Feb 18:32
Picon
Gravatar

SLO-SYN 440-TH125 stepper driver from Superior Electric

Hi, some body has the manual for a SLO-SYN 440-TH125 stepper driver from 
Superior Electric?, i google it without luck and the people of Dana 
Motion said i to old for support

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
Permalink | Reply | 5 comments |
headers
Spiderdab | 8 Feb 17:08
Picon
Favicon

Add New Thread

Hi, finally i'm trying to get my new Keling motors and drivers work,
together with a C35 BoB from cnc4pc.

I wanted to use the charge-pump function (that, if i understood, act as
a watchdog), so i need a signal coming from LinuxCNC every 12,5 kHz, as
written in the manual.
On our wiki i've found this:
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?About_Charge_Pumps

well, i made it work adding the function to the base-thread, and setting
that to 40000ns , which means 12,5kHz as requested. (read the link for
that value to come out)

In the same link is written also that one can add a new thread, if the
base thread period is different from what needed (mine was 49k and
something..). So do you think is better to add a new thread dedicated to
the SCHP (charge-pump), or to lower the base-thread period and work with
that?
If a new thread is the option, how can i add a new one? i've tried with
the real component 'threads', but it doesn't work (maybe my fault).

if it helps, my Tested Base Period doesn't go over 4000ns.

Thanks, Davide.

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
(Continue reading)

Permalink | Reply | 5 comments |
headers
Don Stanley | 8 Feb 08:50
Picon
Gravatar

Is the Intel D510M0 Processor still a good choice for a new EMC2 application?

Hi All;
All my shop EMC2 applications are running on D510M0's.
Is this also a good choice for the new one?

After EMC2 was taught how to deal with the D510M0's printer port strangeness,
the have done great except for boot up after being powered down several days in
a 60 degree F shop. I can't determine it the chips need a little warming or if
the battery needs a little charge.
The fix is easy, just push the reset button when they stall in boot up and they
boot up ready to go.

 Has anyone had similar experience, or know what causing the hang ups.

    Thanks
        Don

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
Permalink | Reply | 14 comments |
headers
Kirk Wallace | 7 Feb 23:53
Gravatar

Updated wiki VFD Modbus

I've added more information and links to the VFD Modbus wiki page. I'd
appreciate any comments, corrections, or advise of copyright
infringement.
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?VFD_Modbus 

--

-- 
Kirk Wallace
http://www.wallacecompany.com/machine_shop/
http://www.wallacecompany.com/E45/index.html
California, USA

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
Permalink | Reply | 0 comments |
headers
Claude Froidevaux | 7 Feb 14:40
Picon
Favicon
Gravatar

pluto FPGA on LPT

Hi,

I'm trying to get a pluto-servo to work. Right now the status is that 
output is OK (led dimming, I test several output ok), but input is not 
working at all, I cannot read anything (input or encoder value). After 
some reading, it seem that I don't do the right choice choosing pluto 
board, as this solution is only marginally working (as I understand at 
least). Is there any hope I can make it to work ? (in a reliable way of 
course !)

my setup is a atom motherboard (D945GCLF), with a additional LPT 
extension (MOSCHIP mcs98651v). So far, I use the not EPP mode 
(ioaddr_hi=-1), I tried most possible configuration without success.

In case I'm doomed, I will most probably try with a mesa 5I25 board. Can 
anyone confirm that it work ? (this time I ask first !!)

Also, if the pluto fpga board is really that much problematic, I thing 
It would be nice to put some warning on the documentation.

Best regards,

Claude Froidevaux

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
(Continue reading)

Permalink | Reply | 8 comments |
headers
Ian Eagland | 7 Feb 12:37
Picon
Gravatar

Re: EMC and 3D Printers was AXIS default jog speed on A-axis (Jon Elson)


Hi

I have been using EMC to drive a 3D printer for over a year. I leave A 
axis as angular. I just make sure calibration is such that 1 degree is 
equal to 1mm of plastic feed. It works a treat. I use skeinforge, sfact 
and slic3r to generate g code. Full details of EMC usage for 3D printers 
here:-
http://objects.reprap.org/wiki/Builders/EMCRepStrap

Forum discussion here:-
http://forums.reprap.org/list.php?155,phorum_session_v5=192%3A915a494e81aad9a9232be59f44d2c878

Code generation discussion here:-
http://forums.reprap.org/index.php?phorum_session_v5=192%3A915a494e81aad9a9232be59f44d2c878
In the Software folder.

--

-- 
Regards

Ian Eagland
Eagland Machine Tools Ltd
http://www.eagland.co.uk/

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
(Continue reading)

Permalink | Reply | 0 comments |
headers
Erik Friesen | 6 Feb 13:32
Favicon
Gravatar

Control board

I am doing a bit of 3 axis cnc routing on pvc.  I would like to set it and
forget it while it runs.  I have been thinking about building, or buying if
available, something down this line.

Parallel breakout for stepper drive.

Microchip pic32 or similar to monitor incoming steps

Encoders on steppers to compare against steps.

Pic set up as usb HID device, with misc IO for jogging, etc.

Amperage monitoring on 110 out relays, controlled with HID or aux parport
io.

I would set up the amperage and stepper monitors on the control board to
flip a bit if things get out of whack, which would be netted to the
machine-on.

Is there something out there that fits the bill already?
------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2
Permalink | Reply | 21 comments |
headers
gene heskett | 6 Feb 07:35

Different way to do offsets

Hi Guys;

I think after a splitting headache from 3 hours of reading docs (too damned 
sweet too according to my sugar meter), I may have a better idea for the 
offsets.

1. Use tholefinder.ngc to locate home dead in the center of that brass tube 
which is nominally -0.2X and +0.1Y from the left front corner of the board.

2. Raise Z out of tubing but I think I did that in (t/b)holefinder.ngc.

3. By hand using LinuxCNC's fixed jog, move machine +0.1000x and -0.2000y, 
and HOME XY.

This will establish a HOME that is -0.1X and 0.0Y from left front corner of 
board.

The idea of the X offset there was to better center the circuit on the 
board.

4. Run Z up and swap drill chuck and probe for engraving bit.

5. Run to 0.2x 0.1 y, and run slowly down till the bit stops with error 
message as it touches the board. Probe hooked up of course.  HOME Z.

6. Use G10 L2 P2 X0.2000 Y0.1000 Z0.000 to establish that G55 is the 
machines operating position image to be used for all XY ops till top of 
board is done, etch, text & drill.

Z will get diddled extensively particularly as the last *.top.drill file
(Continue reading)

Permalink | Reply | 0 comments |

Gmane