Drive shutting down

[Halfnutz]

OK, heres the story. I have a nice homemade gantry router, the build log is here New Gantry Router , mechanically everything is fine, it works great. Ive been using it now for several months.

When VCarve came out, I started having problems with the demand it puts on the Z axis, it would work, but I had to go slow or it would start missing steps. So I bought 3 425 oz/in steppers and hooked them up th a drive system I got off of eBay. Its a commercvially designed system by a now defunct (bought by National Instruments ) company called nuLogic.

The driver has a 44V 6.4A power supply running 3 IMS483 drivers. They are 3.0 amp, 44 volt drivers. The resistors are set for full 3.0 (4.4 V pk) amp operation. So in other words, I am right at the limits of this thing. The motors are 3V 3.0 amp.

What happens is that at certain points, usually the same points, in a VCarve program, or probably any intensive , long program, when all drives are in motion, the Z axis (this is the only motor that gets hot) just shuts down. I am also losing steps at other points, on all axis.

I talked to the guy at IM S and he said it could be the inductance of the motor, I geusse these drivers are limited to motors with less than 10uH of inductance, or the power supply losing voltage at full output.

I have a simple question, how can I run this drive off another power supply? I am confused as to how to run the grounds. The 5 Volts for the logic is at common ground with the main power supply. Can I just hook up the other PS or do I need to ground it to the system also? In other words how do I ground PS #2? The drive has a positive motor voltage input and its ground, and just the 5V input for the logic, coming off an optically isolated seperate supply, but its ground is commmon to the main PS I would imagine.

The other option would be a 44V 10Amp supply, or a diferent driver altogether, but I still need to figure out how to ground the second PS. Im probably making a big deal out of nothing, but any feedback would be appreciated.

Maybe its something entirely different.

[NC Cams]

Ultimately, you'll need to tie all the grounds together or else you'll start to generate ground currents wherever it is possible to do so. When this occurs thru a signal line, you can start to fry low current drivers which gets messy.

Now, when you do set up a ground current, you're running the risk of setting up a ground current from the P'S's with surplus current to the one current hog that is asking for lots of current at that point in time - the thing will try to suck power from whereever it can.

Bigger P/S won't hurt. You might also add more capacitance for extra surge capability.

Where you add the capacitance may be something to play with. Try adding it (large cap) near the servo amp. You can use heavy wire from cap to amp and amp to motor (IE 12 or 14ga) and small wire from P/S to cap (18 ga). This way, you'll deplete faster than you can replenish. You can also monitor voltage to see where you need to add extra OOMPH.

At some point, it should become self apparent if you're gonna need more caps or more P/S capacity to keep up....

A volt meter across the caps will also show if and where voltage is dropping which is clear indication that P'S can't keep up with the demand being asked for at that point.

Like Tim Allen says, "you need more power, grunt groan".

Like racers say, "too much power is ALMOST enough".....

Effectively, the motor with the most load will ask for the most current. Feed it what it wants and it will move/cut at the speed desired. If not, things won't move as desired/intended....

[2muchstuff]

You mentioned that you got 3-425 oz/in. motors, are they the same. If so switch motors with another axis and see if the problem repeats itself. Is the z-axis balanced or need to be. The motor may be working harder one way verus the other way, gravity.

[Halfnutz]

I just found out something is up with this router. I configured it for Mach 3, (Ive been using TurboCNC up until now) and it seemed to be working better, but as soon as I start cutting, everything goes south. I never got it to the point the drive shut down, it was missing steps, just not moving much at all. With the router off, everything is fine. I geuuse Ive run this router to pieces. Maybe it needs brushes? One problem I know I have to fix is unshielded motor wiring. Ive never had a problem with it untill now.

No the Z axis is not ballanced, so yes it definately draws more current than x and y, thats why I thought about running a seperate PS, but now I'm thinking it may just be the router sending back emf to the drive through the unshielded wires.

Now the router is running fine, as long as the controller's not on. Must be simple wiring problem. How simple. Well, it should be easy to fix. That suprises me that the controller is affecting the router, as well as vise versa. When I built this thing I never even thought about wiring. Ive learned so much. My next ones gonna be servo controlled.

[CJL5585]

I do not know how your system is wired, but recommend the following:


The power input terminals of each drive ( + and - from 44 V supply) should have a 1000 uF capacitor rated at 50 Volts DC connected across these terminals. This filters noise and some RFI AND EMI (interference) at the inputs.

You should also have a LARGE filter capacitor (or capacitors) installed at the output of the supply. I would recommend 40000 uF @ 50 Volts DC. You should also have a bleed resistor installed. (The capacitor will absorb most of the inductive currents from the motor windings and reduce some generated EMI from the steppers.)

DO NOT MAKE ANY CONNECTION FROM THE POSITIVE OR MINUS SIDE OF THE 44 VOLT POWER SUPPLY TO CHASSIS GROUND. (If this connection is made, it allows EMI from the 120 VAC router and other external sources such as a vacumn cleaner at a neighbors house, to backfeed thru the equipment ground into the power supply circuitry)

DO NOT MAKE ANY CONNECTION FROM THE POSITIVE OR MINUS SIDE OF THE 5 VOLT LOGIC SUPPLY TO GROUND. (noise and RFI and EMI can get into the logic circuits creating havoc with machine control)

Grounding for proper operation of electronic equipment is different from that of Electrical equipment.

The power supply and logic supply should NOT have any common connections between them. The incoming electrical voltage (assuming 120 VAC) ground should be connected to the electronic equipment chassis and all the equipment enclosures. (This includes the drive housings.) This allows the electronic enclosures to act as Faraday shields protecting the electronics from some lightening, RFI, and radiated electrical noise.

Reading your posts, it appears to me, from years of CNC field service, that your system could have some ground loops, and possibly noise, EMI, and RFI is getting into the power supply and logic voltage supply.

I also think that you might want to add an additional power supply with those steppers. Measure the power supply voltage with all axis moving, and you might see the voltage drop to 25 or 35 volts.


Jerry

[Al_The_Man]

Apart from the regulations that deal with grounding, there is common practice that has evolved over time to reduce problems due to the integration of low voltage electronic equipment and Electrical power equipment.
Some of these practices vary, depending on personall preference.
Some swear by keeping every distinct piece of equipment isolated.
I personally favour the opposite approach. The first thing I do when designing a system, is to set up a cabinet ground buss, which is connected to the incomming power ground conductor.
To this common ground point, I ground all supply commons, both low voltage and high, PC power supply common, shields for all shielded conductors, Including motor DC power supply common and local 120vac control power neutral is also connected.
In any system it is almost impossible to prevent alternative ground paths (loops), but this method assures that the designed ground path is usually the path of lowest resistance for all ground conductors.
In all the years of designing and installing CNC and PC based systems, I have not had a noise interference problem when following this procedure.
Like I said, everyone has his own preferential approach to grounding.
Al.

[Halfnutz]

Well, thanks for both perspectives. It appears as though consistancy is the most important issue. What was confusing me was the negative side of the 5V logic supply. I geusse that the +5V is referenced to its own isolated input voltage, which is common to all power supplies and the chassis. That makes sense. If I add another PS i just need to make sure its input is grounded to the same ground as all other PS's. I think you both (AL and Jerry )are saying the same thing, that all inputs are grounded to the same common ground and chassis. None of the PS's outputs however, are connected to ground, only their input common.

I geusse what I was wondering, is should the negative output of the second PS be connected to the negative of PS #1. I imagine not, only its input should be common, at ground.

Now what about the computer? For some reason running Mach3 exagerates the behavior, and it is clear that noise is being coupled between the CNC system and the router. The computer should be plugged into the same grounded circuit as the drive system, but what about the parrallel port? Are the ground pins of LPT at physical chassis ground? Is there anything to ground at the parrallel port? I can find this out on my own of course, but it makes good discussion.

Man the thing goes nuts now, like it has gotten drastically worse. Allmost like some insulation or shielding in the router is arcing or bridged. Both work fine independantly, but as soon as they are both on the CNC system and the router both go nuts.

[CJL5585]

.
Now what about the computer? For some reason running Mach3 exagerates the behavior, and it is clear that noise is being coupled between the CNC system and the router. The computer should be plugged into the same grounded circuit as the drive system, but what about the parrallel port? Are the ground pins of LPT at physical chassis ground? Is there anything to ground at the parrallel port? I can find this out on my own of course, but it makes good discussion.

Man the thing goes nuts now, like it has gotten drastically worse. Allmost like some insulation or shielding in the router is arcing or bridged. Both work fine independantly, but as soon as they are both on the CNC system and the router both go nuts.

Are the ground pins of LPT at physical chassis ground?

Ans. NO

Is there anything to ground at the parrallel port?

Ans. NO

If you have a connection between the CNC electronics chassis and the ground pins on the breakout board -- then you have a connection (or maybe a short) between the negative side of the 5 Volt supply to chassis. The connection should be found and removed. The 5 Volt supply should not be referenced to chassis ground.

Jerry

[Halfnutz]

Allright! I got it calmed down late last night and this morning. I had mistakenly connected the router into the same outlet as the computer, and the controller into a seperate circuit. I had the cords to thew router and controller switched. It is fine now.

I had no idea a ground loop could cause all that commotion. Thanks for all the help, I was ready to buy a new router untill you guys brought up the grounding subject.

[Al_The_Man]

For what it is worth, I checked all the PC's we have (the non-Laptops), several Dells and Asus, and all have the parallel port grounds at chassis potential and the chassis is connected to incoming supply ground conductor, this is probabally due to the practice of the power supply common obviously connected to the board ground plane, which is in turn mounted to the metal chassis with a ground plane screw, the incoming power ground conductor is secured to power supply case, hence the ground continuity through to the parallel port.
Obviously with a Laptop the ground connection may not exist.
Al.

[NC Cams]

Electrons.

Can't see them, therefore can't trust them.

This should be a must read thread when it comes to grounding and ground loops....

[Halfnutz]

Thanks Al, now I dont have to wonder if mine is unique. My laptop has a floating chassis and LPT shield, but the desktops are all common to ground. I havent tested the ground pins though. The shield to the parallel port cable from my other (HobbyCNC) controller floats, (when disconnected from the computer).

I would assume that it (PP/CNTRLR cable's shield) could be connected to chassis ground at the controller end as well, as long as the computer and controller are on the same line input. If they were on seperate AC circuits it could create a problem.

This all brings up another interesting question, the subject of grounding the router. If I were to add shielded motor cables, they would logically be grounded at the controller end, but what about the gantry frame? I had assumed that the frame was grounded through the router motor housing and mounting bracket, but this is not the case. The metal router housing is isolated from ground, so at this point the gantry frame is ungrounded. I am inclined to think that grounding it through the motor wire shield would be the correct thing to do, but I could be wrong here.

Since it is working good now, I might not change anything, but it would be good to shield the motor cables I'm sure. Weather or not to ground that shielding at one or both ends is what I'm not sure about.

It seams like a big deal over nothing, but wouldn't it be nice to completely eliminate any chance of missing steps and be able to increase feed rates?

Thanks again for the help everyone.

edit-I just learned that the shield should be connected at only one end.