Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: BarryB on January 11, 2015, 06:06:23 PM
-
Hello, I just had a horrible day using the mill. It kept crashing about every 3 minutes. I'm running it on a laptop through a sub smooth stepper, the cnc4pc breakout board, and gecko drive motion controllers. The mill is large about the size of a car, and six axis. The frame is 8020, hsd spindle, b/c head, a table rotor and gantry system. I'd like to make this bulletproof, so I can do real work on it. Right now, I can't. Anybody have steps to build on, test with, so I can depend on this thing? If anybody is in the OC area and would like to lend a hand, I'd love it.
I know scales are all working, and all axis work. Nothing is reliable though.
Wits end, argh!
-
Hi BarryB:
Sounds like an interesting machine, and must be impressive with six axis going. The problem is most likely noise , and with six axis motor cables running around, that makes for a lot of little "antenna's" out there to pick up stuff. Also, a vfd generates a lot of noise as you know.
The usb version of the SS is not as noise resistant as the ethernet version for starters. The general run of the mill USB cables are not very good for shielding, so you should definitely get the highest quality USB cable for the SS. Another simple addition would be to put those "chokes" on all the cables. You can get the clamp on type at Radio Shack, and put them on both ends of cables. Most newer electronics have chokes on the power cables as you may have noticed. Another requirement for a CNC machine is a copper grounding rod driven into the earth, and all the grounds from the shields connected to the copper rod with RG8 shielded cable stripped back to expose the woven shielding; solder copper lugs to the ends, and connect the machine cable shields to the copper rod with those braided type cables. This kind of woven cable is better than a single wire for noise, as it is a function of area, to carry the high frequencies. Do a search for how different machine builders have addressed the noise problem with their motor controllers/VFD's. Really good machine grounding goes a long ways to insure machine stability. Another source of noise is a TIG welding machine. They can emit high frequencies onto the building wiring, even if it is not in your shop. It doesn't take too much to upset the tiny 5v step/direction signals going to the motor amps, or the limit switch signals.
I am not in OC, but am not too far away, but am fairly busy for the next few weeks.
John
-
Thanks John. As you can tell, I'm crazy fed up. I'll go out and get those chokes and a higher quality USB cable. There is a ground going from the frame of the machine to common ground for the house. You mention 'grounds from the shields'. I'm not sure what that means. Hmm, do you mean the woven shielding in a cable? Perhaps that's part of my issue, my motor wire isn't shielded. The spindle wire is, but the motor wire isn't. Perhaps I should rewire this thing on top of this.
Barry
-
I'll go ahead and order more cable and rewire. This is the wire I was thinking of ordering for the motors.
http://www.wireandcabletogo.com/18-AWG-4-Conductor-Communication-Cable-Foil-Shield-500ft-or-1000ft-Spool.html
Think it's up to the task?
I'm asking what are probably dumb questions, but I really don't want to have a mill crash session again. I've gotten to not trust it.
Barry
-
I would try a different brand of drive on one axis and see if that helps.
I had Gecko drives on my first router and suffered the same kinds of problems. Everything had to be perfect electronically for them to be dependable. I'm not slamming the drives as they are very popular, just relaying what worked for me.
-
Can you explain what you mean by the mill crashing.
Hood
-
Hi Barry:
Yes, the shielding on the cables must be connected to a common ground. All the ground wires should terminate at a single, common point.
It would be best if the ground was terminated in a copper rod driven into the ground. You can get such copper grounding rods from Grainger, and also Electrical supply houses. The rods are usually about 4 to 5 feet long, and about 5/8" diameter. The spindle motor cables must absolutely be enclosed within shielded cable, and also connected to the common ground.
I mentioned the RG8 co-ax cable because it has a braided shielding. Strip back the plastic for about 6 or 8 inches each end, and cut back the inner wire. Then solder a lug for a #10 size screw on each end. Use these braided ground cables to link to your copper rod from the other machine cables. The braided cable has more area than a single copper wire, and the hi-frequency noise is conducted away due to the larger surface area of the braided cable. Using the house ground is depending on a single copper wire to go to ground, and not as efficient or effective as the copper rod/braided cable method.
Both of my commercial CNC milling machines have a copper rod just outside the building, close the the machine locaton, with a very large copper cable connecting the machine ground. The have been very problem free.
Another possible source of noise proofing your machine would be to convert all your limit/home switches to 24 volt, and drop to 5v just at the breakout board with opto-couplers. The potential between the signal and ground is nearly 5 times what the 5v signal is, and is much more noise resistant. All commercial machines use 24v signal wiring.
John
-
Hi Barry:
I forgot to say, the wire with shielding you have the link to is excellent quality. Two conductor would be good for limit switch wiring, but 22 ga would be sufficient for limit/home signals, as that is very low current.
John
-
Damnit! I absolutely hate when I see advice to drive a ground rod to solve a noise problem on a machine. If that works it only proves that you have a problem you don't understand and are clueless about how to actual fix it! Sorry for the harsh words but I have seen this kind of bad advice since the first days of PCs, computerized cash registers etc. If you ever get this kind of advice from the manufacturer of any kind of electrical machine you should be very afraid working around it because it very well could run amok and hurt you!
A 90 hr week I'm bit grumpy.
-
Hi garyhlucas:
Sorry you are in a mood to curse my discussion of ground problems, and/or me personally. Electrical noise is present in CNC machines due to many sources; variable frequency drives create hi-frequency noise for one. It is invisible, it is silent, it is elusive, and it can cause problems with computer signals and circuits.
You said "If that works it only proves that you have a problem you don't understand and are clueless about how to actual fix it!"
I would say," If that works it only proves electrical noise was the problem."
AC wiring does have a ground path built into the system, but a direct ground for a machine frame with a shorter path than the building wiring is not going to cause all hell to break loose, and make the CNC machine run amok and hurt people; however you are certainly entitled to your opinion, and your obvious rage.
As far as getting this advice from a manufacturer, I personally know of (4) CNC commercial brand new (within the last year) milling machines and lathes, installed by licensed electrical contractors, that put a copper ground rod through the concrete floor, near the machine for an additional ground, before the machine was even turned on. I suppose you would call them idiots. All of these machines work without problems of any kind; they don't even "run amok!"
In conclusion, I would say, lighten up Gary, were on this forum to offer advice to fellow machinists trying to get their CNC Mach3 driven machines running happily, and without problems; no need to get angry.
;)
John
-
Hood, Mach 3 would freeze which in turn would stop the mill. This would happen within milling for five min. This was super annoying. I'll replace all the motor wires, limit switch wires, use cable, and add chokes on all the cabling. Then I'll test again.
Thanks mrprecise, I'll get 18 awg for the motors and 22 for the limit switches.
Btw, my mill build is on this site. I've always had issues with it, so happy with the help.
Garyhlucas, I won't be putting in a ground rod, but will run a common ground. I'm in a live/wirk loft and don't have the option to drop that outside. I'm hoping it's the shielding that will do it.
Barry
-
John,
No they are not crazy, they are just wrong. Having been an industrial electrical contractor for about twenty years in my first career, and having built control panels for almost 50 years now I have never once found the need to install a ground rod at a machine of any kind. I have seen lots and lots of machines that have had noise problems and when you investigate closely it is never a surprise why that is the case. In case of home built machines I see I am more amazed that they run all rather than whether they have an occasional problem!
Electronic engineers are a big part of the problem. They just don't seem to grasp that the grounding done for shielding purposes is completely different than than the grounding done for safety purposes. Grounding for shielding is all about making sure there is one and only one path from every point to ground. Grounding for safety is all about making sure the voltage difference between any two metal parts on a machine is as small as possible, and you want to create as many paths to ground as you can. In the case of shielding noise your are trying to keep out very small currents at possibly high frequencies. In safety grounding you are trying conduct away what may be many thousands of amperes of current from a fault in motor, heater, etc.
So back to that ground rod. It is driven into the earth. What is the resistance of that earth? Well I can tell you that it is so high that it cannot conduct away a large current surge from from a faulted motor. If the fault lasts the ground will be warm around the rod, and the rod will be at a voltage that will kill you, and yes I have actually seen this happen. I figured it out when I kneeled down to look at something and the concrete was hot! That rod is also a second path to ground, the other is through the safety ground wire that comes in with the power conductors. Ohms law teaches that the current in a circuit is the same everywhere, while the voltage varies with the resistance. So that means that leakage from motors, transformers, solenoids etc. cause all kinds of cray voltages to flow on the ground path, and you have two paths to complicate matters.
So a little checklist for you home builders:
Shielded wire on all sensors that connect to the CNC?
Shielded cables on motors that generate noise servos, steppers, VFD controlled?
All shields connected at only one point on the whole machine?
Ground wires from all electrical control devices installed and connected to just one point?
You checked with an ohm meter to make sure the shield isn't connected to the case of any device with a metal case mounted to the metal frame?
Power wires and control wires routed in separate raceways?
Control wires cross power wires at right angles and don't run parallel?
You are using a PC with a communications cable. Is the shield in that cable tied to chassis of the PC and then connected to ground on the card in the machine? OOPS that is a ground loop!
Let there be chips!
-
You are using a PC with a communications cable. Is the shield in that cable tied to chassis of the PC and then connected to ground on the card in the machine? OOPS that is a ground loop!
Let there be chips!
that's a good one to know I will be changing tomorrow
-
A couple of other thoughts for you home builders:
If your wiring looks like a bowl of spaghetti you are bound to have noise problems
"I'll" neaten it up later" doesn't work, later you find out that some wires are too short and you won't rip them out and lengthen them.
Tie the wires down as you go, use cheap ties and cut them off when you add more wires, finish with good ones.
Once you strip off the shield twist pairs together as far you can. Twisting is an effective shield.
Wires that are not in cables that are pairs (to a limit switch for example) should be twisted when possible.
"Bury the grounds, and neutrals) my dad taught me. Grounds are very important and never change, the same with neutrals. So wire them first so that they are at the bottom of the wiring bundle, not on top of wires you may need to change.
Think about you wiring layout before you mount components. Always try to layout components for the least amount of wire and to minimize wires going in opposite directions in a bundle.
Draw a schematic or wiring diagram before you start. You will be amazed at how quickly you will get confused if you don't, and how many wires there really are!
If want help, and you don't have a drawing you are pretty much on your own, and electricians bill by the hour!
-
It's funny, that when I built the control box I followed all these steps of using shielded cable, keeping high and low power away from each other and crossing only at 90 degrees. I made a visual schematic before I wired it, and then I tied everything down as I went along too. I didn't do that will the mill proper though. I'm not sure what I was thinking back when I wired that portion, although, at least the spindle does have shielded cable. The wire will be delivered this weekend, but I've scheduled a solder party with some buddies, so we can solder all the cables in a couple hours at the end of the month.
Barry