LinkBack URL
About LinkBacks
It's best to set up a star point ground system firstly.
Here is one of a few past links on the subject of grounding.
Serious noise issues
Al.
| | |
Hey All,
I have a CNC lathe that is giving me an issue. I am still not getting consistent movement in the motors and after thinking it was a hardware issue and trying a bunch of stuff, Mariss pointed out that it could be "noise" in the system.
Basically, the system "creeps" during movements. The further I go, the farther off it becomes. I've run the automatic step thing in Mach 3 so many times I've lost count. It comes back to 0 every time though.
I did a google search, and wow...how do I narrow all of this down? Here is what I have:
An old Gateway tower computer case with the CNC controls installed.
Parker OEM750 drivers
C10 Breakout board
Intek (I think that is the name) power supply.
The case is a tall tower case. While it is big, stuff is still sort of cramped in there. I can take a picture of it when I get home and post it.
Wiring is shielded cable. I did not use the ground wire in the cable.
My guess is grounding. I don't have the drivers grounded to the case anywhere. I also don't have the breakout board grounded to the case and not even sure there are provisions to do that with the BOB. Pin 7 on the 750 is a ground of some sort. Should I ground the case to the house through a copper wire to the water pipes and then ground everything else to the case?
Anything else I can do to isolate the issue?
One thread that discussed system noise suggested using an oscilloscope. I don't have one, but may be able to get one and someone that knows how to run it if it would help isolate my issue.
Thanks
Mike
It's best to set up a star point ground system firstly.
Here is one of a few past links on the subject of grounding.
Serious noise issues
Al.
CNC, Mechatronics Integration and Custom Machine Design (Skype Avail).
“Logic will get you from A to B. Imagination will take you everywhere.”
Albert E.
Interesting. This is my first task...get this thing grounded somehow.Originally Posted by Al_The_Man
If I am understanding "star point ground system" correctly, this is just a bolt installed in the metal frame of the case with the ground wires going to it....probably with a star washer for good contact?
A few questions:
1. On my BOB, the motors are hooked up to 2 - com - 3. Do I ground that com terminal to the case too?
2. I read in one of the post to ground the controller end of the shielded cable to the case but leave the motor end disconnected. This does not sound like it is going to do anything.
3. I have two power supplies in the case. I have the 65v one running the drivers/motors and the original computer power supply running some cooling fans. I'm wondering if the computer case ps is giving me issues?
Thanks for replying Al. You've been a huge help on my endeavors.
Mike
It's referred to as Star because all grounds 'radiate' out from one single ground point, the service ground should also terminate there.
The grounded case of the PC should also goes to this point.
Up until recently the practice was to ground one end of shields as there was a concern that grounding both ends would create ground current loops.
The present recommendation is to install what is known as a bonding conductor to each motor, gantry and frame etc, all metallic parts of a machine and terminate at this star point, this is known as equipotential bonding, the recommendation is that if this is carried out, the shields can be connected both ends.
Much of it is covered in this Siemens paper.
http://www.automation.siemens.com/do.../emv_r.pdf?p=1
(Check out page 6.3).
I always ground the power supply commons to this point, but it is up to you what method to adopt.
In any event your PC power supply and M.B. are most likely grounded and common to one side of the AC supply already.
Al.
Last edited by Al_The_Man; 10-18-2011 at 03:30 PM.
CNC, Mechatronics Integration and Custom Machine Design (Skype Avail).
“Logic will get you from A to B. Imagination will take you everywhere.”
Albert E.
Wow. That looks like a good read. I printed the thing out already. I see a couple of areas that I really need to look at.
I know I need to address the grounding issue before I look at anything else.
Thanks again.
Mike
I did some voltage checks on the C10 bob. Checking from common to Dir+, I get 4.95 volts on the carriage and the crosslide axis. When I check the Step+, I get .06 volts on the carriage and .03 on the crosslide.
According to the Parker OEM750 manual, it says this about the voltage for the Step:
The step input is optically isolated. Driving the step input differentially will provide the best noise immunity. Your input
driver must provide a minimum of 6.5 mA—approximately 3.5
VDC. With no external current limiting resistor, the current is
controlled by the applied voltage. This is due to a fixed voltage
drop of 1.7VDC on the opto LED and the internal series
resistor (243Ω). Increased voltage will result in increased
current.
And this about the Dir:
The input is optically isolated. It may be differentially driven.
Your input driver must provide a minimum of 8mA at 3.5VDC
to ensure proper operation. With no external current limiting
resistor, the current is controlled by the applied voltage. This
is due to a fixed voltage drop of 1.5VDC on the opto LED and
the internal series resistor (243Ω).
Is this an issue??? Evidently I am OK on the Dir (minimum of appoximately 3.5VDC), but the Step seems to be a little short.
Why do I have almost 5v on the Dir, but almost nothing on the Step?
Update: Figured it out. Had to check the step low active in the ports and pins in Mach 3.
Mike
Last edited by BrassBuilder; 10-18-2011 at 10:19 PM.
The P.P. common should already be at ground through the M.B. and P.S. being grounded internally.
Al.
CNC, Mechatronics Integration and Custom Machine Design (Skype Avail).
“Logic will get you from A to B. Imagination will take you everywhere.”
Albert E.
I am taking P.P. is the parallel port cable going to the BOB?
So...in reference to my BOB voltage question above, I had to turn "on" (green check mark) the Step Low Active and leave "off" (red X) the Dir Low Active in the motor outputs in Mach 3. Is this normal? Any time I've seen examples of this screen, both are either green check marked or red X. I haven't seen where it was set up one of each.
If I turn on the Dir Low Active, then I don't have volts going to that side of the BOB.
Mike
The setting of Dir as low active (or not) may not make any difference other than reversing the motor direction with respect to the value of the Dir signal.
Are you using a multimeter to measure the voltages?
Can you jog a motor in different directions while monitoring the voltage on the corresponding Dir terminal? It should be close to 0 in one direction ("logic low") and close to 5 volts in the other direction ("logic high"). The Dir signal is steady as long as the motor should be turning in a particular direction, and then as the direction changes, the Dir signal will change to be steady but with the opposite value.
As for the voltage on the Step terminal, the Step signal is a pulse, rather than a steady signal. Or, to be more clear, the voltage on the Step terminal will be steady, either at a logic "high" voltage or a logic "low" voltage (depending on your Mach3 setup), when no Step pulse is being transmitted, and then very briefly it will pulse to the other logic level and then back again when the step pulse is transmitted. But you will not be able to view such a pulse on a typical multimeter. Rather, the multimeter's voltage reading will just tell you the "steady" state value.
So, when you set the Step signal to "active low", that means that the "steady" state value is "high", and then for each step pulse it rapidly switches to "low" and then rapidly switches back to "high". By "rapidly", I mean that the step pulse may only be 5 or 10 microseconds in duration, which is fast enough that you would need to use an oscilloscope or logic probe to detect it, but you could not see it on a multimeter. That's why you didn't get a voltage reading on the Step terminal until you set up that signal with Mach3 to be "active low", because with that setting the steady state is "high", which you can read on the meter. If you set that signal to "active high", then the steady state would have been "low", which would show up as no (or almost no) voltage on the meter.
I was using a multimeter. I don't have an oscilloscope and wouldn't know how to read one or work one if I had it. The Dir had close to 5v moving the motor one direction and 0 going the other direction. That seems like it is working as it should. This would also explain why I was not picking up voltage on the Step side.
Mike
I'm in the middle of rewiring my control box. My Pacific Scientific motors have a terminal labeled ground. Do I:
1. Hook up the ground wire on my shielded cable to this terminal? (edit: probably not after doing more reading...and I guess the correct term for that wire would be a drain wire.)
or
2. Run a ground wire from the motor terminal over to the motor mount or the lathe?
or
3. Do nothing with it?
Thanks
Mike
Last edited by BrassBuilder; 10-27-2011 at 10:06 PM.
WoooooooHooooooo!
I GOT IT!! Where the heck are those dancing bananas???
Attached is a picture of my control panel and I'll do my best to explain what I changed.
1. Grounded the system using a star ground. Most of the grounds terminate at this point. The exception being the power cables feeding the power supply and the KB controller. Those are grounded at the bottom of the case. I grounded one end of the shield here and also ran a wire from the ground port on the Pacific Scientific motor to this location. The wire I used was 18/4. Since I did not have a dedicated wire for ground (I was leary of using the shield wire for that purpose) I ran another 18 gauge wire alongside of the shielded cable and taped it every few inches. More on this later.
2. I moved the breakout board from near the bottom of the case to the top. I originally had it mounted right in front of the KB controller. I now have all the data controls at the top third of the case and the power supplies on the bottom two thirds.
3. I added two ferite snap choke cores onto the two motor lines coming into the case. No idea if these actually help, but I was going to use anything I could to cut down on noise. The package had "Helps prevent EMI and RFI." Figured it couldn't hurt.
4. I originally had the motor cables hooked up to (2) 9-pin d-sub connectors. I did away with them and just ran the motor cable as one continuous line. I figured it was one less connection in the line and less chance of another problem. I would really like to put the connectors back in sometime.
5. I searched for "ground" in all my manuals. One interesting tidbit I found was that Compumotor recommended a ground run from the OEM750 housing to the case. Done.
The wiring looks sort of like a hodge podge...but trust me....it is better than what it was.
I did continuity checks on everything that was soldered to a pin. I found three connectors (2 on one motor and 1 on the other motor) that were flaky. From what I understand of continuity checks with a multimeter, the 3 digits on the display should be fairly stable. The three problem ones bounced all over and wouldn't settle on one number. I resoldered those three and now everything checks out good.
I also mounted my Compumotor drivers on a heatsink. Before I just had them mounted on a 1/4" thick piece of aluminum.
While browsing around ebay, I found shielded wire as shown in the second picture. It consists of two 18 guage wires in a shield with ground wire, a second pair of shielded cable with its own ground, a separate ground wire and then a foil wrap around the whole thing. I have 100' of it coming. I planned on wiring the step ports with one set, the dir ports with the second and then run the ground from motor to case. It should be a cleaner set up.
So....what did I gain?
Before the system would creep the farther the axis moved. If I got it tuned to move 3", if I did a 6" move, it would move 6.005. Same way with smaller movements. If I moved .010, .020, .030, .040, I would get readings of .011, .021, .032, .043....always a couple to a few thousandths off.
Now...it holds steady at less than .0005 no matter what I do.
Thanks for all the help!!!
Mike
Posting Permissions
| | |
