[Fish4Fun]

John,

It has come to my attention (crash) that installing limit switches on my A4 is something I should do sooner rather than later :drowning: LOL. I thought I would ask your advice before I started, seeing as thus far you have not only built a great machine, but have given excellent advice to go with it! The concept is quite simple, and this may well be like "making extra tables" where I am over thinking something simple, but I thought I would ask FIRST for your input this time :-)

I made my first "real part" on Tuesday! I needed a part for my minnow tank, and while I could certainly have done it with my manual router table (or manual mill, or perhaps even a hand drill and jig saw, lol ) I designed it using V-Carve and then cut it out in foam to "test" the design. After a tweak for closer fit, I cut one more time in foam, and then cut it out of a "cutting board" (a $3 cutting board turned out to be the right thickness and have all the properties I was looking for, LOL). While the part fit perfectly, elapsed time from start to finish was ~ 6 hours. I could have easily done a similar job w/o CNC in less than an hour, but a lot of the time was wasted attempting to draw the part in Sketch-Up (and then Solid Works) and then import the drawing into V-Carve. Finally I simply drew the part in V-Carve (actually easier than Sketch-Up or Solid Works) and from there it went pretty quickly. So, the 6 hours was essentially 5 "lost hours" to "learning curve", and 1 hour of actual "work on the project" LOL. Next part I will start in V-Carve, and will likely take considerably less time.

I am still waiting on some "V" router bits to arrive (China is a LONG WAY AWAY!) I have 10 * 60 degree and 5 * 45 degree bits coming (total 15 bits), all with 0.1mm tip diameters. I have a fairly extensive assortment of PCB end mills ranging from 1/8in down to what looks to be (the smaller ones aren't marked, lol, any thoughts on accurately measuring them?) ~0.5mm, with a fair number in the 1mm to 2.5mm range. As I mentioned previously, I am hoping to do some PCBs. From reading, it appears Eagle with the PCB-gcode.ulp script is "easy" to learn/use, but I am investigating DXP Designer, which, in theory, can do CAM output. Any thoughts you have would be much appreciated!

Fish

[microcarve]

Hi Fish,



Yeah, VCarve Pro saves me tons of time too. I doubt there's any 2D part that
I couldn't make in a pretty quick time...much faster than *any* software I've
tried...and I tried a lot before I bought it.....

Once you get even a little practice using it, things get even easier. There
are all sort of things it does to make absolute placement of holes, sizes, offsets,
etc...easy-simple. Best software I've ever heard of or seen.

I still to this day haven't bothered with limit switches on these small machines.
But I Do make a habit of test running new files ---a few times---to see how
they'll run. 2-3 times is usually plenty to see if it'll be reliable.

But someone else may want or need them. Especially when using the machine at it's
extreme limits of travel....which I rarely do, but the times I have, I've actually
had pretty good results from setting up Mach3's soft limits. It took me a few times
to get the idea of how it'd act, but it's not hard to make work well enough to
avoid all the wiring that comes with adding physical switches.

Also, cnc machining will create the finest dust possible in my experience. As fine
as the finest baking flour. That stuff gets everywhere, so I'd always been
a little concerned about the switches lifetime usefulness....though in all fairness,
I never gave it a real chance.

If I did, I might consider some that Mactec54 had recently pointed
out on another thread....(I can't find it now that I'm looking...but they were
a blue colored proximity switch...I think for $10 each....very high quality)
I figure if that's his choice, it'd be impossible to go wrong with them...with
the supremely high quality machines he makes....

Ah!...here's the place....

Pricing

Whether those are suitable....maybe someone can add their thoughts...but
they look like some ideal switches --best I know.

I use a lot of different home positions due to different fixtures that I bolt
in place for various parts, so my home is always different as well. So I like to
just use these little knobs from Radio Shack to make very fine positioning
easy just by cutting off controller power to "unlock" the motors, then re-zeroing
the new position....

Hexagonal Control Knob - RadioShack.com

That's always been easy for me...but I did develop some bad habits early
on. I will say that trying to jog a few thou...on Z in particular...has cost me some
nice carbide bits a couple times. Those fine little sharp edge tips are easy to crack
off....

I do like those blue switches though. I don't think the price is at all unreasonable
for what they are and how much trouble they'd save someone.
I haven't had luck ever with low budget switches.

The motors will stall harmlessly at any rate if they go too far. They'll make
a helluva noise, but it's harmless.


I never had time to do any PCB milling myself, but I have concentrated on
getting the machines to the point where they'll be able to do it as well as it
can be done on any machine in it's class. My little ultra fine lithographs use
a very very fine stepover and it can be seen under high magnification that
it does an excellent job consistently. So positioning will work fine. It's then
a matter of practice with getting the board flat & level and figuring out
the right feeds/speeds/depth & cutting bit. That will take some trial & error,
but it'll work great...

I am looking forward to seeing how PCB milling goes on the machines.
I have some ideas for milling fixtures, so I'm waiting to see how well
it's working before I make a few of those. I wish I had time to experiment
with it myself, but I've sure thought about the details of it for a long time.

The Eagle software...I haven't gotten anywhere with yet. Time...


John

[Fish4Fun]

John,

Thanks! I did have a crash that actually broke the Y-Axis nut loose from the PVC container. Super-Glue fixed it, but I THOUGHT I might need something smarter/faster than the panic button, LOL. You are 100% right about, "If I hadn't run it w/o trying to make the part (and set things up right), it wouldn't have happened", but still I am thinking about installing them to protect me from my silly mistakes, LOL.

John, honestly I am just beginning to learn the machine. The software has been much harder than I ever imagined, and the Vectric venue is BY FAR the best I have played with, but to do PCBs, I will OBVIOUSLY need to venture elsewhere. I am currently playing with Eagle and DXP. I FAR prefer DXP to Eagle from a schematic capture point-of-view (simple circuit with ~400 LEDs took close to two hours w/Eagle, 10 minutes w/DXP). I have done it in both, defining the PCB was easier in Eagle, everything else was easier in DXP. I have NOT generated gcode on either platform...yet, so the "upfront time" may prove equitable with the "back end" included; but I do believe that the $$$$ for the DXP will produce viable results.....I hope.

I have been VERY remiss in posting pictures of my build. I want to post some pictures of my cabinet, but I just now have received the t-nuts to finish the cabinet build, and wanted it to be complete, so expect some pictures this week.

When I get PCB MFG "right", I will post the results with accolades, I am 100% confident in your machine, I am 10% confident in my ability to make the software do "what it is suppose to". Anyway, I could NOT BE HAPPIER with my purchase.

THANKS!

Fish

[mcphill]

I heartily endorse the Eagle/PCBGcode combo. Eagle is a little "awkward", but once you get a hang of it, it is not too bad. There are huge downloadable libraries of components, and you can define your own as well (again, it is odd the way it is done, and I have to find the instructable every time to relearn it, but it is fully functional). I have been VERY happy with how easy using PCBGcode is, and you should also run Opti.exe on the output to get rid of a ton of "air cutting" that is normally generated.

If you want to see some of my results, from the first time I played with it, go here: PCB Work | mcphill | Fotki.com, photo and video sharing made easy.

[microcarve]

One of the fixtures I keep thinking about for PCB
milling goes like this....

A slab of, say, 8" x 8" x 3/8 or 1/2" aluminum plate
with a slab of a 1" thick cast urethane plate attached
to it.

There are different kinds of urethanes that are so
slightly different to look at them, they appear the same,
but the differences can matter.

In this case we'd want one that's very hard and highly chemical
resistant.

The thick aluminum plate is for being able to bolt it into place
repeatedly and without distorting the top urethane plate. For
repeated and long term use, really...

The fixture, bolted into place would then be "planed" or leveled
in relation to the bridge/gantry.

If a fairly wide bit is used, the toolmarks will tell if the
Router/Z/bit is exactly straight up & down....which will be
pretty necessary for making very fine traces. It'll be adjusted
at that point and the plate re-leveled.

The reason for the urethane plate is that it's much more easily
leveled than if it were aluminum and at the same time, it's
highly chemical resistant.

The urethane can be drilled into with no harm to a bit.

(It'd be considered a 'consumable')

Chemical resistant, because the PCB can be super glued with
only a very little bit of CA glue....just only what'd be
absolutely needed to reliably hold it in place for the job.

Then a few drops of Acetone will allow it to easily be released.

A wall or pattern/pocket could be milled in initially as a
"registration" point for the circuit board....so succeeding
ones would fit exactly in the same place.

I haven't actually made one of these things yet, but it's a
recurring idea that gets sketched out pretty often, and I'm
not finding much fault with the concept.

I'd see it more useful for making several of the same circuit
board at a run...as it being removed would possibly need to do
another resurfacing of the urethane. But I think it may
have some good potential. Maybe for other things like inlays
as well. Any reason to hold very fine thinner materials in
place for milling.

I've been meaning to make some drawings, but it's not hard to
imagine...

Once I get some time, I plan to experiment with one....


John

[Jay C]

On the subject of home switches. I use them and it's for one reason. If i'm in the middle of a job and there is a stall, I can home, and restart and it'll go right to where it needs to. John, you say you set your "home" all over. But really you are setting 0,0 which you can do with home switches.

Mach has many coordinate systems. The machine coordinate system is absolute and measured from the home switches. I have a [nomedia="http://www.youtube.com/watch?v=wJJzIv5gaqc"]YouTube - Cutting an enclosure for the SuperPID[/nomedia] video that shows what I'm talking about. The machine stalls twice but I was able to recover the job.

I use simple opto interrupters (milled the circuits on my machine) and have never had issues with dust fouling them. Then again I never abuse my machine under a pile of shavings like you showed with the Alumilite :P

FWIW,
Jay

[microcarve]

On the subject of home switches. I use them and it's for one reason. If i'm in the middle of a job and there is a stall, I can home, and restart and it'll go right to where it needs to. John, you say you set your "home" all over. But really you are setting 0,0 which you can do with home switches.

Mach has many coordinate systems. The machine coordinate system is absolute and measured from the home switches. I have a YouTube - Cutting an enclosure for the SuperPID video that shows what I'm talking about. The machine stalls twice but I was able to recover the job.

I use simple opto interrupters (milled the circuits on my machine) and have never had issues with dust fouling them. Then again I never abuse my machine under a pile of shavings like you showed with the Alumilite :P

FWIW,
Jay

Well I wouldn't want anyone to think my machines look like that
all the time....

That was purposely done to show that Oiltes will tolerate quite
a bit. It only took a few hours to get it that full too....but it
does stay clean. Vacuum is right next to it 24/7...

I can't recall having a job stall in years. I've only tried to pick up
a position once, and it was more trouble than it was worth. An
HDPE nut had worn out on me so it was a lost cause anyway.


John

[ger21]

If i'm in the middle of a job and there is a stall, I can home, and restart and it'll go right to where it needs to.

If your machine stalls once, then you're trying to go faster than it's capable of. A properly designed stepper machine should NEVER stall. Now, if you're trying to get every last bit of performance out of it and can live with the stalls, that's up to you.

But I agree with you 100% about being able to re-home and start without having to setup anything. You can start a job one day, turn off the machine, and finish the next day with no aggravation. Or, if you have to hit E-Stop, just re-home and continue on your way.

[microcarve]

Here's some simple opto switches I made a long time ago...
just to see if blue LEDs I had on hand would work.

Flickr: kz1927's Photostream

They did...and very well, but with the machine contained...and
the fine dust that MDF cutting generated, they would get "off"
a bit from time to time.

It was also an experiment. I like to push the machines to see how
much they'll take and still work reliably from time to time. Reasoning
behind that is, I can fix anything that goes wrong with one easily here,
but someone else may have a hard time. Knowing what could cause trouble
in advance has been useful....

Naturally, my opto's could be much better designed to tolerate problems,
but then....I'm constantly changing things, so it's good to know how much
abuse many things will stand up to.

R&D....

[microcarve]

Well, I've got pictures of my table fixtures, but I can't find them...

Anyways, What I do is fix material to a separate piece of MDF.
I may have 2-3-4 same size pieces that have material affixed
for quicker loading or some other reason.

Those "fixtures", due to variances in sizes of them for any
particular part, or mostly because of a fixed pattern of t-nuts.
may be placed in different places on the machine table/bed.

The files were also generated using different software over
a period of time and for widely varying machine sizes. So it's
not practical to recreate the files to match whatever machine they're
used on at the time, but it's a very simple matter to relocate
X0Y0 to a mark on each of those fixtures and use that as "Home"
without any physical switch.

The "mark" on each of the fixtures is most always just a hole
that I move X & Y to. It can be off a few to several thousanths and
still run fine.

So basically, each one of these fixtures has it's own Home
position to cut whatever's affixed to it. They can be run on
any size machine that's here just by locating that "Home"...X0Y0
hole in each table.

Where the fixture is doesn't matter. Whatever t-nut pattern makes it
simpler to bolt to varies with each size....or machine.

They're generally run for a day or so, so relocating X0Y0...which
is what I'm referring to as "home" wouldn't work with a physical
switch.

Limits...I haven't had any need for on these small machines, I just
make sure to test run each file first.

If that makes any sense at all....

It's worked for years on many, many machines.


John

[Fish4Fun]

mcphill,

Thanks for the link & advice! I looked @ some of your other pictures including the 1610L, that looks like a really cool machine!

I am curious if you have a preference in bits for PCB engraving? I have some Chinese bits on the way, mostly because they were cheap and I figured I would break a fair number during the learning curve, but once I am done breaking bits, I would like to invest in some better quality ones, I would love to hear your thoughts.

Additionally I am curious if I should be thinking about a "roughing" pass with a medium size (1/32nd to 1/8th inch) "fish tail" bit prior to doing the detail work with the V-Bit? Or perhaps it is just best to leave copper "islands" on the PCB? I have not made it all the way through either the DXP or Eagle renderings (I am still working on part placement...a real pain with 400 LEDs, LOL).

Thanks in advance for any other tips you might have!

John, yes, I have been thinking on a PCB work table. I like your ideas and will continue to think/tinker with the table until my new bits come in. I have been thinking on an aluminum surface for a while (pretty much since before the machine arrived), but the idea of epoxy/urethane for a sacrificial top is very intriguing. I have a couple small medical surplus vacuum pumps that I might be able to mount on the table itself and create "jig plates" for the various size PCB material and hold them down with vacuum....that would be really cool! I will have to dig up those pumps and review their specs (I have a whole box of them a friend sent me several years ago, don't remember why he even had them, lol).

As always, lots of good thoughts!

Fish

[mcphill]

I use V bits from Drewtronics, I prefer as steep an angle as possible so that minimal depth changes don't have huge effects on the width of the cut. I like the 30 degree tips here:

DrewTronics

I also use a bunch of different tiny end mills. I get them at Midwest Circuit Technology: Router Bits | Diamond Files | Carbide Drill Bits and Carbide End Mills | Carbide Tooling - Midwest Circuit Technology

I got the drill bit sets, and quite a number of different mill diameters. I posted in my Fotki what diameters I used. I do a roughing pass with as large a bit that can be effective as I can, then go to a smaller bit to clean that up. If I have VERY small traces, I finish it off with a V bit.

Sometimes I change all that up and cut with the V bit first, then "hog out" the rest with a pretty small bit.

I am no pro - just trying different strategies to see what works for me.

[jig123]

John,
Here's few pictures of my progress so far.
A picture of my electronics box with the Gecko g540 mounted. I put jacks in the case to connect the various sensors/control signals: i.e. e-stop, x/y/z-min-limit, spindle on, etc. Got tired of (un-)screwing the connections from barrier strips. So far, seems to work fine.

A picture of my first attempt at milling plexiglass/acrylic to try to determine speed/feed. All done with a 2 flute 1/8" end mill, most at 15k rpm. All of the top row attempts were/should be on the same horizontal axis. First/left-most was at 22ipm w/ .07"DOC and as can be seen, its not exactly a straight cut. I believe the feed was too fast and also the bit/spindle was far enough "extended" from the z gantry carriage to cause significant lever arm movement. Next cut was 15ipm w/ .07"DOC. Still too fast for that arrangement. Then 10ipm/.03doc. That appears to work mostly ok. Then 10ipm/.06doc, again not good. So I'm currently using 8-10ipm/0.03doc.

Then I moved on to trying squares. With these, even that speed/feed caused some movement out of true horizontal/vertical. I decided I needed to move the workpiece closer to the z-gantry (as you have mentioned in many posts!).

But I made my first attempt at using the gcode from a heeksCNC/CAD model for a touch plate I want to make. (3rd picture).
This was done at 15k rpm, 1/8" 2 flute end mill, 0.03doc, 0.06 overlap. Looks somewhat ok, but the center cylinder cut has quite a bit of melting/junk. The 120degree channels look ok. I think I need to try a slower spindle speed. I'm guessing the toolpath spent too much time near the center/bottom of the cylinder and heated the plexi.

To build up the work table and provide a better way of holding things, I took a piece of 2x10 and mounted it on top of a 3/4 piece of plywood on top of the 3/4 MDF on the y carriage. I moved the spindle up accordingly. I milled the table top to make it flat to the x/y and then milled channels in the table to hold some T-rack aluminum. That's the last picture.

Now I need to go back to trial plexi and see what speed/feed looks better and then redo the touch plate. And finally get to mill channels in my vacuum table (HDPE). To hold a PCB. So many projects, so much enjoyment.... ;-)

[Walky]

I am curious if you have a preference in bits for PCB engraving? I have some Chinese bits on the way, mostly because they were cheap and I figured I would break a fair number during the learning curve, but once I am done breaking bits, I would like to invest in some better quality ones, I would love to hear your thoughts.

If you got some of those 10º chinese bits (the ones with one flat side), be careful about DOCs and feeds because the 0.1mm tip snaps too easily (it might have been cased by router's runout, though, but I'm not entirely sure). They'll probably still be usable for non-SMD work anyway. I've seen some chinese "pyramid" engraving bits that I gues must be a bit more forgiving (gotta get some of those sometime).

Here's a test I made with one of the 10º bits (very crudely grinded into a radiused bit after the tip snapped):

[nomedia="http://www.youtube.com/watch?v=oM2acy0VPrA&feature=channel_video_title"]YouTube - Grabado de PCB en fresadora CNC[/nomedia]

BTW, those bits love fast feeds.

[microcarve]

John,
Here's few pictures of my progress so far.
A picture of my electronics box with the Gecko g540 mounted. I put jacks in the case to connect the various sensors/control signals: i.e. e-stop, x/y/z-min-limit, spindle on, etc. Got tired of (un-)screwing the connections from barrier strips. So far, seems to work fine.

A picture of my first attempt at milling plexiglass/acrylic to try to determine speed/feed. All done with a 2 flute 1/8" end mill, most at 15k rpm. All of the top row attempts were/should be on the same horizontal axis. First/left-most was at 22ipm w/ .07"DOC and as can be seen, its not exactly a straight cut. I believe the feed was too fast and also the bit/spindle was far enough "extended" from the z gantry carriage to cause significant lever arm movement. Next cut was 15ipm w/ .07"DOC. Still too fast for that arrangement. Then 10ipm/.03doc. That appears to work mostly ok. Then 10ipm/.06doc, again not good. So I'm currently using 8-10ipm/0.03doc.

Then I moved on to trying squares. With these, even that speed/feed caused some movement out of true horizontal/vertical. I decided I needed to move the workpiece closer to the z-gantry (as you have mentioned in many posts!).

But I made my first attempt at using the gcode from a heeksCNC/CAD model for a touch plate I want to make. (3rd picture).
This was done at 15k rpm, 1/8" 2 flute end mill, 0.03doc, 0.06 overlap. Looks somewhat ok, but the center cylinder cut has quite a bit of melting/junk. The 120degree channels look ok. I think I need to try a slower spindle speed. I'm guessing the toolpath spent too much time near the center/bottom of the cylinder and heated the plexi.

To build up the work table and provide a better way of holding things, I took a piece of 2x10 and mounted it on top of a 3/4 piece of plywood on top of the 3/4 MDF on the y carriage. I moved the spindle up accordingly. I milled the table top to make it flat to the x/y and then milled channels in the table to hold some T-rack aluminum. That's the last picture.

Now I need to go back to trial plexi and see what speed/feed looks better and then redo the touch plate. And finally get to mill channels in my vacuum table (HDPE). To hold a PCB. So many projects, so much enjoyment.... ;-)

I haven't had time to do any acrylic milling, but best I understand it,
the cast type works far better...and IIRC...the protective film on it
can sometimes indicate whether it's cast or extruded.

Extruded is supposedly a problem with melting. But... I see tons of
stuff with a Google search where guys who make acrylic computer
cases mill it all seemingly fine. And with nothing more than a Dremel
in some instances.

I think the brown paper film on the plex means it's cast...but I could be wrong.

I have several sheets of various plastics I intend to try when/if I can
ever find time....

The guys who make the speed controller called Super PID have some
videos of a router slowed down and doing *Excellent* acrylic milling.
Seems like it was in the 5k speed range.

www.SuperPID.com - Super-PID Closed-loop Router Speed Controller

I need to get one of those things.


John

[Walky]

The guys who make the speed controller called Super PID have some
videos of a router slowed down and doing *Excellent* acrylic milling.
Seems like it was in the 5k speed range.

www.SuperPID.com - Super-PID Closed-loop Router Speed Controller

I need to get one of those things.

Get one, they're awesome! there's no way I'm going back to the infernal full-RPM noise of my router (for me, the noise factor is easily half of the SPID's value).

[microcarve]

Get one, they're awesome! there's no way I'm going back to the infernal full-RPM noise of my router (for me, the noise factor is easily half of the SPID's value).

I'm becoming more and more convinced.

And more...and...


John

[Mick40]

John,

Where did you purchase the epoxy in bottles? I searched and only found it in tubes.

Excellent thread!!!!

Thanks,
Mick

[microcarve]

John,

Where did you purchase the epoxy in bottles? I searched and only found it in tubes.

Excellent thread!!!!

Thanks,
Mick

Thanks Mick!



Epoxy 330

If you're just doing coloring of letters like I did, then any 2 hr epoxy
from most any hardware store will do fine.

I avoid 5 minute epoxies when I can, they have gone soft and flexible
after a few years. Problem is you just never know if it'll stay hard for
a long time.

Epoxy 330 is the stuff they put in class rings in some case, so it's been
very reliable.

Rio Grande also has some fancier, more expensive epoxy kits that are
really good stuff. It can be polished...which most epoxies of these
types can't.

For most things, though...common easy to get 2 hr type works very well.


John

[Mick40]

Hey John, thanks a lot! I have been following your build (s) and I'm impressed, your very talented. Appreciate you taking the time to share it means a lot. So much to learn and your thread has helped me more than you know.

Much appreciated,

Your pal,
Mick