One more question:
The G540 manual shows a simple estop circuit which uses a NC switch to ground pin 10 (enable). http://www.geckodrive.com/upload/G54...4%20MANUAL.pdf
There is another document of the Gecko site showing a latching relay type estop which cuts power to the PS and has a load dump resistor. http://www.geckodrive.com/upload/Ele...nicalEStop.pdf
Is the latching relay prefered? is the load dump resistor necessary?
Will,
Very interesting thread, thanks for sharing. I've got two of these plates kicking around , one with some test holes drilled into it, though in my case I was thinking of a mill at the time.
Regarding long bolts loosening over time, it could happen due to different thermal growths of the materials if the machine is somewhere it will thermal cycle a lot, such as an uninsulated shop/garage. Loctite or locknuts should be cheap insurance against that.
Regards,
Raymond
Use the G540 wiring. The other is for servo motors.Originally Posted by will gilmore
On the previous page (http://www.geckodrive.com/support.aspx?q=10003) it mentions using that design for stepper or servo.
Is having the estop open the connection between the enable pin and ground a robust enough estop stystem? or should I have the estop cut power to the power supply?
As I understand it, the E Stop using the enable pin will halt the controller s/ware outputting any further movement commands so it works fine for the majority of people but if the motor wiring spazzed out and the controller internally faulted to allow power through uncontrolled then breaking the AC power line would be a good idea........just got to weigh up the likelyhood of that occuring with a Gecko?
fair enough. I'll see if my estop switch has two sets of contacts otherwise I'll Just go for the enable pin.
I'm still trying to figure out the following:
Do I need a 5A fast blow fuse between PS and drive? (out of curiosity why is a fuse ok when a switch is not ok on DC side of PS)
Do I need 470uf 100v cap across PS inputs on G540?
Do I need zener diode voltage clamp across PS inputs on G540?
The above are in reference to http://www.geckodrive.com/upload/Step_motor_basics.pdf pages 10-12.
What gauge wire should I be using. I've looked on the internet and there are many charts saying many things. Is there a standard for wiring inside an enclosure? and for general machine tool wiring?
I'm planning on using 4 conduction 22 gauge shielded wire for the steppers and inputs. I'm looking for suggestions on the following:
Wires from PS to G540 (48VDC,8.3A max): ?
What size wire for all the 5VDC circuits(estop, inputs, outputs etc)?
AC to PS: 120V, 400W = 3.33A?
Also, any one have a good idea/source for connectors for inputs? db9 like G540? DIN? ??
D-Sub DB9 connectors are easily available at any electronics store. Digi Key will also have them.
I am using five pin XLR'S on my next build, as the axis have a little more amperage than D-Sub can handle.
Wiring size for signal wires? Kind of irrelevant concern. If you think of the minute traces on a computor motheroard surviving these voltage and amperages, any wire you choose as being big enough to solder/crimp easily, and have enough mechanic strength for the moving bits will be a overkill electrically.
I have used .75mm flexible, shielded cable on my lathe for the last few years and it has worked fine.
Regards,
Mark
Sorry forgot to answer you question about the wire size from PSU to Gecko. I used some scrape wire I savaged from something with is probably about 2.5mm sq.
I didn't bother to look it up, as it looked like an overkill.
AC to PSU, I think I butchered an old lamp or extension cord cable.
Regards,
Mark
I've been learning about the electronics needed to run this thing. I'll do a full report when I get it all together. I got a couple 2U rack mount cases from a friend. Took apart a perfectly good computer and reassembled it in one of the case. Hope it still works. The other case has the G540 and the power supply.
I've been avoiding aligning the head stock to the ways and sure enough I'm having some issues with it. Here is what I've done so far:
1) put the upper Z rail with one bearing on the surface plate. use dial test indicator to get it aligned with the edge of the surface plate.
2) put lower Z rail on surface plate and get it aligned to first rail
3) put ground steel plate on Z bearings, use dti to get it square to Z rails
4) put the head stock on with 1/2" drill rod in collet and roughly align to Z rails
5) using dti on drill rod about 1" from collet, rotate spindle and set dti 0 in middle of dti swing.
6) repeat step 5 at 3.75", 6.5", 9.25", 11". turn nuts on threaded rod to adjust headstock alignment. the goal is to have the runout at each point remain centered at 0 on the dti.
here is where I am now(don't give me grief about significant figures I'm just estimating):
1": 0 (<.0005" runout)
3.75": +.001" (~.003" runnout)(plus meaning swing of dti is centered farther up)
6.5": +.001375" (swing from -.0005" to +.00325", .00375" runout)
9.25":+.000625" (swing from -.00375" to + 5", .00875" runout)
11": -.001" (swing from -.008" to +.006", .014" runnout)
I went back and swept the edge of the granite with the dti now that I have the ground plate and all the bearings on there to see if anything has changed. The whole travel is within .0005". The first 8" or so are pretty much right on and then a little bit off at the end.
wow, long post. I can't believe you are still reading.
I guess the conclusion I'm drawing is that either the edge of the granite plate is curved and therefore my rails are curved or the 1/2" drill rod is drooping. Anyone see a problem with my method? Should I get a thicker test bar? I have up to 1" collets no chuck yet. Should I just stick the 1/2" out 6" and assume the rest will line up? Any advice or encouragement is welcome.
Last edited by will gilmore; 02-20-2010 at 02:34 PM.
Because the problem is when switching the power on. A fuse can only kill it.(out of curiosity why is a fuse ok when a switch is not ok on DC side of PS)
Gerry
UCCNC 2017 Screenset
[URL]http://www.thecncwoodworker.com/2017.html[/URL]
Mach3 2010 Screenset
[URL]http://www.thecncwoodworker.com/2010.html[/URL]
JointCAM - CNC Dovetails & Box Joints
[URL]http://www.g-forcecnc.com/jointcam.html[/URL]
(Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)
I think you could be seeing both problems. I don't think there is any guarantee that a surface plate is straight on the sides.
I also calculate that your beam is drooping
Deflection of a beam due to distributed load = W * L^4 / (3 * E * I)
Where W is the load per length, L is the length, E is the modulus of elasticity, I is the moment of inertia
11" of 1/2" steel rod:
Area = PI*r^2 = 0.196 in^2
Density of steel = 0.284 lb/in^3
Weight of rod = Area * density = 5.576e-2 lb / in
Moment of inertia of the rod = pi*r^4/4 = 3.068e-3 in^4
Modulus of elasticity of steel = 30e6 lb/in2
deflection = 5.576 lb/in * 14641 in^4 / (3 * 30e6 lb/in^2 * 3.068e-3 in^4) = 2.96 e-3 in =0.0030 in
You aren't measuring that beam droop. You are measuring on the slant of your slant bed lathe. Let's say it's 60 degrees from horizontal. Then I think a rough conversion is that you are seeing 1 - cos(60) = 50% of the above number or 0.0015", which is pretty close to what you are really seeing.
As a goofy idea, have you seen how much deflection is coming from pushing on that bar with the indicator? You could just push another indicator on the bar and see if it moves enough to care about.
I'm loving this build. I think you are going to have a cool project here.
Last edited by ebrewste; 02-20-2010 at 12:56 PM. Reason: misnamed variables
Ger - ah ha! makes sense. unfortunately I put the PS right next to the G540 to keep the wires short. do I a) run all the way to the front panel for a fuse, b) run a fuse close to the G540 which will require taking the enclosure out of the lathe if it blows, or c) do nothing. I believe I read that the G540 has an internal fuse.
ebrewste - all good points. I believe drill rod is HSS and bed is at 45 degrees. Don't know about HSS vs steel but at 45 degrees I'll see slightly less droop than you calculated (30% vs 50%). I'll check out the indicator pressure. I only have one dti and only one mag base. I can hold my plunger dial indicator against the bar with my hand or put the dti on the other side.
I'm also measuring with a spherical tipped dti on a cylindrical rod so ...?
So where do I go from here?
Always fun to check in on this topic. Beautiful build!
Different steel alloys don't change deflection or weight much, so drill rod vs mild is a moot point.
The 45 degrees slant is a good point and you calculated the difference right.
The single indicator is tough. Without knowing the indicator contact pressure, you can't calculate it. I don't have any problem with holding another indicator by hand, but it needs to have similar force. Try pushing each against your finger. If you can use the stiffer one (dial indicator?) by hand and see no significant deflection, you eliminate the lighter one as a problem. No need for another holder.
The cylindrical rod and spherical tip don't matter if the deflection is small compared to the rod diameter. I would say 1-2 thou vs 1/2 rod is a small effect.
You could also tip the whole thing back so you have 0 degree slant to eliminate drooping as a culprit...
I plotted the deflection, and it looks pretty "droopy": http://spreadsheets.google.com/ccc?k...YUFNTWEE&hl=en
I didn't go and calculate the theoretical droop to compare. It might be a good idea to tilt the machine and try to eliminate the droop as the culprit. Or pull out the strength of materials text.
that means a lot coming from the builder of the most beautiful lathe around!
ebrewste -
Thanks for putting your brain to work on this. When I go to the shop I'll check out the indictor pressure first. Don't you think putting the dti on the other side of the bar and comparing measurements would work?
Tipping the machine is easier said than done. ~130 lbs of granite plus a bunch of steel and the headstock. It's pretty top heavy. I guess I could weld on some extra legs to hold it so the granite plate is level...
Before I started this tootalew proposed getting an 18x24 plate and splitting it into two 9x24 plates to save weight. If I had done that I would have a nice surface to check the straightness of the rails with. nuts. oh well I guess weight is good in the long run.
-Will
Regarding indicator pressure on test bar: short story is I don't think it has much effect with my little DTI. Long story: with the DTI against the unsupported end of the test bar I pushed my dial plunger indicator against the test bar. Least deflection I could get was .002" and went up from there as the plunger spring compressed. uh oh. then I took the DTI off the test bar and pushed the plunger directly against the needle of the DTI. I couldn't get the plunger indicator to move at all until I used all of the travel of the DTI. So the DTI has much less pressure than the plunger indicator.
After that I messed around with the headstock a little more and then decided to move on. Its pretty easy to adjust, if I'm getting measurable taper I'll go back.
I think this is very sensible. The pictures look great!
Good that it is adjustable. After a crash you will have to realign it anyway.
Regards,
Mark
I checked my calcs and the deflection is only 3/8 as much as I said. I don't think it matters in the end.
I also updated my spreadsheet with a simulated deflection. I also made a simulation that was "skewed" (misaligned headstock). It shows that most of the non-straightness is just non-straightness. I'm not sure where it comes from in your system.
The equation I put below says:
Deflection of a beam due to distributed load = W * L^4 / (3 * E * I)
Where it should be:
Deflection of a beam due to distributed load = W * L^4 / (8 * E * I)
That'll teach me to get a value from the internet and not derive it myself.
Last edited by ebrewste; 02-22-2010 at 11:08 PM. Reason: Added equations
progress?
Having some computer/mach/gecko issues but Mariss himself responded to my thread (http://cnczone.com/forums/showthread.php?t=100710) so I'm confident it will be resolved.
Will
Reply with Quote
