PDA

View Full Version : ? for resident homebuilt CNC gurus



Rustybolt
01-31-2008, 07:02 PM
Finally got my Seig2 (mill that fell off a truck) mini mill together and wired. The problem is that while the steppers turn they don't turn consistantly. Sometimes the steppers will just stutter and sometimes they will do what they are supposed to. just as often if I jog in a specific direction the steppers will do just the opposite.
Curriculum Vitae.

Mach 3, Gecko 201s, campbell breakout board. keling 23 steppers. Motors tuned as per Mach 3 manual(however both dir and step pulse freq. are maxed out at 5.)

Right now I'm thinkin' there may be interference somewhere along the line. But I'd like to hear from a pro.

John Stevenson
01-31-2008, 07:13 PM
Sounds like a problem with the 5v logic to the drives.
Not familar with the Campbell board though as I use the CNC4PC boards and these have an outside 5v source.

Keep power cables to the drives, motor cables and logic cables seperate.
Check grounds.

.

.

Evan
01-31-2008, 08:05 PM
Shields on data cables , motor cables and power supplies should only be grounded at one end, never both ends. Data grounds run inside the shield. Grounds should be in a star configuration meaning that all parts of the system including the shields should share the same ground point. Neutral on the mains should NOT be connected to ground in any of the boxes.

Sparky_NY
02-01-2008, 08:36 AM
Data lines that are inputs, limit switches for example generally cause false switch trips in Mach3 if you have noise/grounding problems. (boy did I learn that the hard way).

In the motor tuning, the times must be at 5us for Gecko 201's or they will appear to run properly but actually miss many steps. You said you are set at that but I figured I would mention it anyways.

Few things to check/verify:

With power off, do things turn nice and free by hand?

Gecko 201 current limiting. What is the current rating for the motors you are using? What value resistor do you have on the Geckos?

A very very common mistake, that causes symptoms like you have- switched step and dir lines!!! Also check Mach3 setup for proper active low/high settings.

Double check motor tuning. Can't tell you what steps/per should be set at approx. without knowing if your direct driving the screws or belt ratio and your TPI of the leadscrew. Give us that info and your present settings and we can tell you if your in the ballpark.

What sort of voltage are you running to the steppers? Is the supply voltage sagging with the motors running?

If the wiring between the Gecko and the motor is mixed up, the motors will grunt, groan, try to turn but just not go. Same for a disconnected wire. Verify the motor winding to gecko hookup. (power off of course!)

There is a few things to check for starters.

(check the step/dir lines are not reversed first and the active high/low, that is a extremely common mistake/problem)

I did a setup almost exactly like yours recently and am now in the process of doing my bridgeport with Mach, gecko's and steppers which should be done in the next couple days. I have made a lot of mistakes and learned a lot along the way.

lwalker
02-01-2008, 01:55 PM
Sounds like you're losing a phase. Check that all the connections between motor & driver are tight (you probably already did this multiple times), but double check that the wires are in the right order.

The behavior you describe is very common when one motor phase is wrong: with no load it may step correctly, under a load or at a different speed it can stutter, or go backwards.

BobWarfield
02-01-2008, 04:44 PM
LOL

Everything these guys have suggested gives the symptom you see and I have seen every one of them in my own CNC work:

- Don't forget to check Mach 3's parameters to make sure they all agree with what you'd like. You can vary whether Mach things signals should be hi or lo to reflect "on" for example.

- I had 3 steppers, but on 1, I reversed 2 wires in the cable, Step/Dir. I knew I had reversed them and thought I could keep it straight. Worked well first time. Then I took it all apart to make a change and forgot step/dir were reversed when I put it back. Got exactly your symptom and took a bit to remember what I'd done. DOH! Now I put labels on the wires instead of relying on the colors.

- Noise, yup. I only use shielded wire now, and try to make sure I get the grounds right on everything.

- I had a "many wired" stepper one time and got one of the coils misconnected. It acted much as you describe.

The good news is that it doesn't take too long to recheck all the wires and maybe conduct an experiment or two.

Cheers,

BW

Sparky_NY
02-01-2008, 05:51 PM
Another thing. It would probably be easiest to just play with one axis at a time. Once you find the issue and get that axis working, all bets are its the same problem with the other axis. Playing with them all at once just makes for confusion.

(Wiring steppers this afternoon, so I know how easy it is to get confused.)

Rustybolt
02-02-2008, 09:02 AM
Thank you everyone. This place is the best. The motor cables are shielded but not grounded. I'll give that and the other stuff a try this weekend.

Evan
02-02-2008, 09:23 AM
An ungrounded shield on a cable carrying strong signals such as motor power will act as a transmitter of noise. Be sure to ground them well at the supply end. If you need to shield a cable or wire set you can buy stainless steel braided shielding from the auto parts performance section. Another source is stainless steel shielded hose in the plumbing department. That makes a good flex cable. Another trick for a shielded flex cable is a screen door spring.

Sparky_NY
02-02-2008, 10:00 AM
Thank you everyone. This place is the best. The motor cables are shielded but not grounded. I'll give that and the other stuff a try this weekend.

Remember, only connect the shield at the controller end, the shield at the motor end is trimmed back and unconnected. Tie all the shields at the controller end to the same point, chassis ground, not power supply ground.

Double check the wiring from controller to motor, a miswired phase is a more likely cause of your issues and your going to be in there anyways.

Got my bridgeport running 100 ipm reliably last night during my initial tests! Today is setting up the limit switches and coolant/spindle relays and I am done! The only other item is to figure out a chip enclosure around the table to keep the mess in the garage down. 3 sides are straight forward, its the rear side towards the column I have to figure out something that contains the chips but clears the ram and column. Anyone with pictures of something would be a real help!

Evan
02-03-2008, 08:59 AM
I though I would pass on a very recent experience with my machine. A couple of days ago I noticed that the X an Y axes were very occasionally hiccupping while idle and/or jogging. The motor would either lose a step or step when not commanded. This was very intermittent and might happen only once in a slow jog across the entire travel. I have been troubleshooting this on and off the last two days and last night found the problem, sort of. A few days ago I removed an extra parallel port card that adds two more parallel ports and put it in another machine to use with Mach 3.

This was the only thing that coincided with the problem. Also, all signs pointed to the computer as the source of the problem. If I disconnected the parallel cable everthing stayed quiet. Now, removing an auxillary port card should have no effect on the on board parallel port. I checked everything and it was all to spec.

Then, I turned off the computer once when the steppers were fully armed to move. The steppers went nuts. This absolutely confirmed that the issue was a noise problem on the parallel port of the computer since the I/O board I am using already has pullups on the buffer inputs. The cable wasn't suspect as it is an industrial quality Amphenol cable with all metal construction and double copper braid shielding.

One way to deal with noise on a data circuit is to load the circuit to lower the effective impedance. This can be done with either pull up or pull down resistors. Since I didn't want to rewire the controller just to troubleshoot I added a small breakout "box" which is a special connector that has a LED for every signal wire in the cable. These LEDs are powered by the actual signals by drawing a couple of milliamps of current from each signal line. This loads the line slightly and will have the effect of reducing the signal levels slightly.

If the noise signal is just at the threshold of being able to cause a state change in the input buffer of the I/O card even a slight reduction will put it below that threshold. The same degree of reduction to the real signals won't affect operation since they have a lot more noise margin.

Adding the mini breakout box instantly eliminated the problem. I still don't have a good reason to why removing another PCI parallel card from the computer suddenly increased the noise level on the on board parallel port. It shouldn't make the slightest difference.

Sparky_NY
02-03-2008, 11:16 AM
I though I would pass on a very recent experience with my machine. A couple of days ago I noticed that the X an Y axes were very occasionally hiccupping while idle and/or jogging. The motor would either lose a step or step when not commanded. This was very intermittent and might happen only once in a slow jog across the entire travel. I have been troubleshooting this on and off the last two days and last night found the problem, sort of. A few days ago I removed an extra parallel port card that adds two more parallel ports and put it in another machine to use with Mach 3.

This was the only thing that coincided with the problem. Also, all signs pointed to the computer as the source of the problem. If I disconnected the parallel cable everthing stayed quiet. Now, removing an auxillary port card should have no effect on the on board parallel port. I checked everything and it was all to spec.

Then, I turned off the computer once when the steppers were fully armed to move. The steppers went nuts. This absolutely confirmed that the issue was a noise problem on the parallel port of the computer since the I/O board I am using already has pullups on the buffer inputs. The cable wasn't suspect as it is an industrial quality Amphenol cable with all metal construction and double copper braid shielding.

One way to deal with noise on a data circuit is to load the circuit to lower the effective impedance. This can be done with either pull up or pull down resistors. Since I didn't want to rewire the controller just to troubleshoot I added a small breakout "box" which is a special connector that has a LED for every signal wire in the cable. These LEDs are powered by the actual signals by drawing a couple of milliamps of current from each signal line. This loads the line slightly and will have the effect of reducing the signal levels slightly.

If the noise signal is just at the threshold of being able to cause a state change in the input buffer of the I/O card even a slight reduction will put it below that threshold. The same degree of reduction to the real signals won't affect operation since they have a lot more noise margin.

Adding the mini breakout box instantly eliminated the problem. I still don't have a good reason to why removing another PCI parallel card from the computer suddenly increased the noise level on the on board parallel port. It shouldn't make the slightest difference.

Very strange indeed. Guess I am not the only one to get these wierd gremlins that pop up. Also very interesting that shutting the computer off increases the noise. My mach3 setup, with gecko 201 stepper drivers, sets perfectly quiet when the computer is turned off. With your symptoms, I'd say the computer isn't the source of the noise (verified by turning it off) but rather the "antenna" to assist in radiating it. Look at the computer as a antenna for the transmitter. (the transmitter/noise generator being unknown at the moment). Removing the board and its cable changes the tuning and impedance of the antenna and thus its efficiency at radiating.

Again, very strange symptoms indeed. Your obviously not a beginner in these things so my comments may be obvious. I'd question if the computer and its case is grounded? I have seen those cheapie plug strips go intermittent too many times or just a bad cord contact, perhaps on the ground contact? Ten to one odds its 60 cycle noise, from differing ground potentials between the computer/controller/mill. With the computer off, there are no squarre waves floating around to cause noise.

My original Mach setup with a small x/y table had a lot of noise issues. I ended up putting .1's on the various steady state inputs and outputs such as limit switch lines and that helped a lot.

By all means, report back on what you discover! It will be a good one to store away in the reference book. LOL

ptjw7uk
02-03-2008, 11:28 AM
Electrical noise is a very strange beast and sometimes difficult to track down and sometimes impossible to stop due to poor design.
I have a strange one at the moment in the workshop with my radio that I have playing in the background just to kil the silence of stationary machines, in that if I use a hacksaw on some oblect in the engineers vice I get static crackles on the radio I just assume I am being a better aerial than the radio which has no earth connection as mains lead only has 2 wires also its just a cheap clock radio.
So it just goes to show that we are surrounded by the air waves which must cause some induced current which could be picked up by the cnc electronics.
Peter

j king
02-03-2008, 10:19 PM
This grounded cables has me interested.I am doing a BP cnc pc conversion and I am replacing the steppers so am going to run new wires since the old ones will be rerouted and are stiff.I was going to buy just a standard multi wire cable but from what I am reading this is not the correct thing to do right?What kind of wire are you using to go to the steppers and limit switches?
Thanks jim

Sparky_NY
02-03-2008, 11:04 PM
This grounded cables has me interested.I am doing a BP cnc pc conversion and I am replacing the steppers so am going to run new wires since the old ones will be rerouted and are stiff.I was going to buy just a standard multi wire cable but from what I am reading this is not the correct thing to do right?What kind of wire are you using to go to the steppers and limit switches?
Thanks jim

Sheilded cable for both. The sheild keeps the signals from radiating and getting into other wires and causing problems. The shield in the cable only connects to ground at the controller end, the motor end is trimmed back and left disconnected.

Evan
02-04-2008, 05:44 AM
Also, all the components of the CNC system should be connected to the same outlet on the same mains power circuit. The computer, controller and any other CNC related and powered devices should share the same gound point. This also applies to any computer setup. Accessories such as a printer, the monitor, amplifiers for speakers etc should all run from the same circuit and outlet.

John Stevenson
02-04-2008, 06:08 AM
This grounded cables has me interested.I am doing a BP cnc pc conversion and I am replacing the steppers so am going to run new wires since the old ones will be rerouted and are stiff.I was going to buy just a standard multi wire cable but from what I am reading this is not the correct thing to do right?What kind of wire are you using to go to the steppers and limit switches?
Thanks jim

Jim,
This will generate some flack but all I use is standard 4 core cable, no shielding and terminated at both ends into plastic potting boxes that now act as terminal boxes.

I do take a lot of care how the components are laid out, cable runs and grounding but that's all. So far I have personally built about 20 machines and have about the same number of working machines out there built from the X3 kit we make.

So far we have had no problems and of the 20 or so I have built about 10 are all different types, makes etc from small tabletop routers to large 3 tonne vertical mills.

On build at present are a Denford Triac conversion from the old North Eastern Electronics to modern motors, controller and drives and the same with a Bridgeport MDI mill.

I must admit I have never had a lot of success with the Gecko drives on large machines.

.

Evan
02-04-2008, 06:25 AM
John,

It may not give any trouble with the control to use unshielded cables but it may interfere with other nearby equipment. The motor drive signals are steep square waves and will generate a long series of strong even numbered harmonics of the driving frequency. How bad this will be also depends on the type of microstepping used. The greater the number of microsteps the more the drive signal approximates a sine wave and the less the radiated interference.

[added]

If using 4 wire flat ribbon cable the phases should be wired to use adjacent pairs in the cable. Phase A should have a pair that are side by side and likewise phase B. They shouldn't be interleaved.

Sparky_NY
02-04-2008, 10:49 AM
John,

It may not give any trouble with the control to use unshielded cables but it may interfere with other nearby equipment. The motor drive signals are steep square waves and will generate a long series of strong even numbered harmonics of the driving frequency. How bad this will be also depends on the type of microstepping used. The greater the number of microsteps the more the drive signal approximates a sine wave and the less the radiated interference.

[added]

If using 4 wire flat ribbon cable the phases should be wired to use adjacent pairs in the cable. Phase A should have a pair that are side by side and likewise phase B. They shouldn't be interleaved.

I fully agree. Having come from a background in radio transmitting equipment, I know only too well how well square waves radiate. Ever notice the FCC approval sticker on computer power supplies and other electronic equipment that have switching power supplies? Those square waves make a great transmitter and without attention to shielding and such they will intefere with radio/tv reception for a long ways. In the cnc case, the normal problem is they get picked up on the signal lines, such as limit switch wiring, and carried back into the controller/computer where they cause false pulses. Stepper motor outputs, are outputs so the lack of shielding wouldn't directly effect the motor operation itself, its their radiating and getting picked up on other wiring in the vicinity that poses the potential problem.

Also, here is another thought. Do you shield the stepper wiring to prevent it radiating or use standard cable and use shielded cable for the signal wiring (limits etc.) to prevent that wiring from picking up stray noise? The norm is to shield both. Using unshielding wiring on the steppers may work fine for the mill but I pretty much guarantee it will raise hell with radio reception in the area, especially AM radio. (remember the FCC approval on switching power supplies?)

We are talking 50 volt or so square waves at 25khz or there abouts, combined with 4 foot or more long unshielded motor wires, we have the makings of one fantastic transmitter! As Evan pointed out, square waves have even harmonics that extend far into the RF frequency range with pretty high amplitude. To cause trouble, the frequency isn't constant, it varies with stepper speed, so what we have is essentially the same thing commonly known as a "Frequency Jammer" similar to what the military used to use to jam radio communications. The cnc mill might not care but radio/tv equipment in the area certainly will.

For John that has used unshielded cabling on many retrofits: You can get away with that on retrofits.... new equipment would never get FCC approval as required.

BobWarfield
02-04-2008, 11:08 AM
Noise is one of those "why take the chance?" problems. Lots of folks don't use shielded cable and they are fine. But if you do discover a noise problem, it can be intermittent and hard to diagnose.

I doubt if the original issue for the thread is noise, but if you're setting out to build a system, it's very easy to come by shielded cable:

http://www.action-electronics.com/wirecable.htm

Look at "multiconductor shielded" on that link. I bought a roll and it'll last a lifetime of fooling around with CNC.

For lower voltages, such as encoders, you can also use CAT5 networking cable. It has its own form of noise protection, but you need to connect the grounds properly.

Cheers,

BW

John Stevenson
02-04-2008, 01:20 PM
Also, here is another thought. Do you shield the stepper wiring to prevent it radiating or use standard cable and use shielded cable for the signal wiring (limits etc.) to prevent that wiring from picking up stray noise? The norm is to shield both. Using unshielding wiring on the steppers may work fine for the mill but I pretty much guarantee it will raise hell with radio reception in the area, especially AM radio. (remember the FCC approval on switching power supplies?)



AM radio ??
I think they shot the last one in about '69

It's all digital now over here.

I can see what you are saying but to be honest I have had no problems.
I have run the small router in the house and it doesn't affect the TV although that's cable and we have the radio on in the shop except when the TIG welder is on and the big CNC but that's because the 440v output invertor plays up. The mill can do dry runs fine.

Next door is closer to the workshop than we are, they are nearly attached but we are 40 foot away and they don't report any problems on interference at all, even the TIG, and they would tell us, they are very good neighbours.

.

Sparky_NY
02-04-2008, 03:08 PM
AM radio ??
I think they shot the last one in about '69

It's all digital now over here.

.

Hopefully they don't have airplanes over there..... LOL
AM was and still is the standard for aircraft. Amazing they never changed it over the years.

Rustybolt
02-10-2008, 09:09 PM
Still no joy. everything is grounded and resistored. Gonna try a different computer this week.