Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Leeway on May 05, 2007, 07:39:47 PM
-
Hey guys.
Thought I would ask this here too. Perhaps someone has an idea that may help.
I am having a lot more trouble with my second build, which is a plasma cutter.
First was a router with a 3 axis Xylotex. I didn't even need a BOB for that one.
I am using Gecko 201's. I had initially setup to use a standard no frills BOB, but just swapped out to a Mini I/O2 board. I had been running the 5 volts from my lab supply, but this mini has it onboard.
I have my step and dir hooked up correctly I think. I have the estop working. All motors are working as well. The Z axis seems pretty stable, but X and Y are intermittent when jogging in either direction. Both directions will go both ways.
I am running shielded cables and they are grounded at the drive cabinet. They run all the way to the motors, but are not connected.
My motors aren't connected to anything. Not even grounded. Just sitting on the floor. I wonder if grounding them would help now that I think about it.
Anyhow.
I am getting the proper voltages at all locations. This mini I/O has the little led card that shows step and dir for the axes when running. The step leds work as they should I think, but the dir leds only come on sometimes when I jog one direction, not the other. That is to say using one key and not the other. The motors still go both ways. Also the A axis isn't even connected and those led's work at times as well.
Now I have the 5 volts connected to the gecko's 5v common. The PC ground terminals on the board are not connected, because I don't know where it would go. Instructions are kinda odd on that. It says PCGND is common for all drives. The 5V terminal says +5 for common to all drives. I realize that the +5 common is labeled that way on the Gecko's, so that was easy. The ground is confusing. I'm pretty sure it should not go to my high voltage ground. That is separate.
Sooooo.......any ideas on what the heck is going on? ;)
Here are a couple of photo's so you should be able to see the wiring maybe.
Thanks.
http://leestyron.com/cnc/cb2.jpg
http://leestyron.com/cnc/cb3.jpg
-
The LED for Dir should be on one way and off when you jog the other, is this what you get?
I am not too clear on what you mean when you say "Both directions will go both ways." are you meaning if you jog using the left button it may go left then next time right?
Hood
Oh BTW think probably the PC ground will just be the ground from your Parallel port but not 100% sure
-
Okay.
Sounds like the red LED is working like it should then.
You are correct about the X and Y driving either direction with one arrow key, then the same when I try the other arrow key.
I have tuned the motors from about 40 Vel and 5 acc to 500 vel and 45 acc and the results were the same. They run pretty good, but errantly.
I have tried the different pulse ranges from 0 to 5 on all axis as well and that didn't effect the stable Z drive. I have not swapped the motors yet with this board, but I will try that.
I have tried the motor drive step and dir active low and they work best on active high.
I have another new cable that will run my THC 300 and second PPort and I can swap that out. I will check it also for straight pass though on the wiring.
I am running this one at a freq of 45000 and its pretty stable. I will try these things, but I am slowly running out of safe ideas to check. ;)
I wonder if I should try to hook the drives up to PPort 2. That may give me another thing to test or eliminate.
Thanks Hood.
-
Sounds like swapping out hardware will be the best move, try swapping the Z drive for the other axis drives and see if the problem moves around.
You will get there in the end and it will likely be something simple, most of us have been there ;)
-
Thanks Hood.
I did swap some stuff out.
I swapped out computers to this laptop.
I removed the charge pump plug in and jumpered the board on pin 1 and 2 like the manual states.
I disconnected the led's.
I swapped out the motors.
All three run well on Z.
I swapped out the wiring between the drives and the BOB.
Tried motor tuning with the different setups.
Swapped out Pport cables.
Checked the cables.
They are straight pass through.
The only thing I have left to do that I can think of is to swap the gecko's.
I will remove the Z gecko and place the Y one where the Z was.
If it runs good there, then there has to be some noise somewhere.
If not, then the two drives will need work.
All three power led's in the gecko's work and stay on.
When I drive the axis, the led's flash dimly.
Not having used gecko's before, I assume that is a normal condition, especially since the stable drive does the same.
I remeber checking the jumper settings in the gecko's, but will double check those.
Next post will be my results of swapping out the drives.
-
Okay.
I moved the errant Y gecko into the spot where the stable Z gecko was. It now works as it should, so apparently, the drives are okay.
Thats good.
I guess its a noise thing.
If you look at my first picture, you can see that the motor wires are above the two offending axes, but not over the third. They are somewhat offset. They are not shielded right there either. Only to where they enter the cabinet.
Could this be the trouble spot?
The shield grounds are also above these drives.
-
Could well be, I have shielded cables on everything in my mill, only wires that are not shielded are the power wires into the geckos.
Hood
-
Do you think an aluminum plate slid in between the drives and the incoming wires would help prevent any noise? What about lead or brass? I have sheets of all three. I would really hate to start all over again on the cabinet. The reason there are no shields in the cabinet is because I used pigtails so I would have plugs and receptacles at the case.
Now if I take it a step further. Can I wrap a ground wire around each drive wire bundle and expect that to help?
I'm not an electrician or electronics guru. I did take basic electrical classes and was a Generator mechanic in the Army, but that was long ago. I do understand the priciples of basic stuff, but I usually deal with noise by changing the channel or turning down the volume. ;D
I don't recall installing a noise reducing knob either, so I am boxing shadows now. Thanks for any suggestions on grounding or reducing noise problems.
-
I think you can buy sleeving to shield your wires, its kind of like a braided cover. Might be worth looking into that. Dont think the Alu or lead plates will help but as I know less about electronics/electrics than you I would take that with a pinch of salt.
Hood
-
I found some that I will try out. It's part number 5537K27 at Mcmaster Carr. Page number 773. $1.41 per foot.
Here is the description.
Braided Tinned Copper Mesh Sleeving
Flexible yet sturdy, this braided sleeving protects wire, cable, and hose in highly abrasive environments. Just cut with scissors; sleeving expands for installation and can be pulled tight around the bundle for a secure fit. Designed for EMI shielding and grounding applications. Color is silver. Maximum temperature is 257° F. Maximum continuous length is 1000 ft.
I think if anything will do it, this will.
Maybe I won't have to rebuild the whole thing. Thanks, Hood.
-
yep thats the stuff, might be an idea to get heatshrink for the ends as well if you dont already have it.
Hood
-
Hey Lee,
A couple of other things that may help or you might want to check. Your connections on the geckos. The blue connectors push on to pins. You can pull them off and rotate 90 degrees and push back on. I have seen some blue connectors that had a bad connection. Also, twisting, (braiding) your wires may help and if you have to cross another wire, run it across as close to 90 degrees as possible And try not to run them parallel to one another.
Brett
-
Nope dont put the connectors on the other way, I got told off from Mariss for doing that ;) Seems they are only meant for the way in the pic.
Hood
EDIT
just been talking with a redneck and he says he had bad connectors so give them a go round the 90degrees like he said, anythings worth a try especially if its free
-
I would contact Gecko and ask for new connectors. They are real good to work with and will most likley get them right out. A good bunch to work with.
Brett
-
Thanks a bunch, Guys.
I did check all the blue connectors. They seem to be fine for continuity. If I turn these around, they don't lock in and the cover doesn't fit right. They don't plug in as far either.
I suspect that if swapping the connectors out has fixed someone's issue, they may have had a noise issue and the movement and different placement of the wires cleared up the issues with noise.
At any rate, just moving a bad drive to a good drives location and using the good drives wiring stopped the issue it was having. It purred just as the good one did.
I don't think it's that actual connections. All the wiring is nearly a short as possible and pretty much identical.
I think the noise issue has to be the trouble. If it were a ground loop, the the stable drive wouldn't work right either.
I wasn't even aware that the blue connectors came off before this though. Looking at them closely, they are a nice little bit of diseign and engineering as well. Double sliding dovetails. Very cool. ;)
Thanks Guys.
-
Well, Guys, even after the shielding, I still have the problem.
Z runs perfect. The other two run pretty good, but will go both directions using the same arrow key.
Here is a photo of the shielding in place.
http://leestyron.com/cnc/cb4.jpg
My site is down for the past few minutes but should show up soon.
Now I am nearly at an end of the battle with this case and configuration. I have no clue what might be going on now.
I have swapped everything I could think of including computers, keyboards, drives, motors, cables, BOB's, 5 volt sources, switches, PPort cables etc.
I have swapped pin settings on the BOB and matched them up to Mach. All three motors have the same settings and then different settings from very slow to fast. No difference. I still get bidirectional movement from one arrow key. I get this same kind of movement regardless of which arrow key. Both on X and Y.
I have a question. The 470 uf caps. Gecko says to use these if the PS is more than a foot from the drives. They are right at a foot. Would it hurt anything to remove them and give it a whirl? I am thinking it would be okay. I am still just boxing shadows n this.
This is all before I even get the THC 300a turned on and the plasma going. This could really take some time. :o
Thanks for any suggestions at this point.
-
I did remove the caps and there was no change.
I am officially out of things to try.
The only things I didn't shield are the 12 volts to the fans from the computer PS. I did shield those switch wires though.
I didn't shield the 120 v ac running to the BOB and the switch either, but I was getting the same type thing with just a standard BOB and using the 5vdc from the lab supply, so I am kinda ruling out interference from those.
They are also running in the very front of the case and not really crossing any of the drive and step/dir wiring.
Just really odd stuff here. For me anyway.
Perhaps its time to try a completely different case.
-
Do you have the shielding grounded? I see some of it is but cant tell if the rest is? Also I think it is best to have all the shielding grounded to the same point but maybe I am wrong there.
Hood
-
Now you see there! I hadn't even thought about that.
The shielding that was originally grounded through the cabeling is grounded. The rest really isn't. Perhaps intermittent at the back where I have the clamp on the three motor plug recepticles, but not actually screwed down for sure.
I will try that and thanks a bunch.
-
Try and ground at one point, think its what they call star grounding.
Hood
-
No joy on the grounding of the shield wires.
I used the brass bar at the rear to act as a grounding buss bar. I then used some of this shielding braid as grounding wire. Attached terminals to one end and gator clips to the other. All shields are grounded to the driver case, computer case and welding cart its all sitting on. I even shielded the ribbon cable that goes to the step/dir led's and also tried it with that disconnected.
I think it will require some more head scratching. ;)
-
I know you have swapped drives about but have you tried swapping the actual wiring, ie take the connector from the Z and place it on the X.
Also might be worth trying moving the wiring on the breakout for each axis.
Hood
-
I can't remember if it has been mentioned before or not. Have you tuned your Geckos? Also, what is your debounce set at?
Brett
-
Thanks Guys.
I have turned the pots on the gecko's to the limit in both directions with no real effect. I have set the debounce all the way up to 2000 and still get the same results. The computer I am using is running at 45000 freq. and is pretty stable.
I haven't switched out the drive wiring yet. I have considered it, but these are all just short new straight pieces of wire running from either the BOB or the fuses.
I was thinking my next step might be to pull everything out of this case and spread it out. Use some longer wires for everything and see if that doesn't help.
It just has to be something simple. Thanks again. I will follow up with my results.
-
I was not meaning changing the wires, I was actually meaning just unplugging from say the Z Gecko and placing the plug on the X gecko and vice versa. This will let you know if its motor or wire problems as the problem will swap.
Hood
-
another thing, have a look and see that enhanced pulseing is enabled.
Hood
-
Ah. I think you mean the blue plugs? I haven't actually swapped those out between drives, but I have mounted each of the drives where the Z is on the far right and they all function well in that position. All three motors work great and as they should when plugged into anu drive in the Z location as well. This is telling me that it has got to be noise of some kind.
I have been focusing on the rear of the cabinet. I have 62 volts coming into the front of the cabinet and all of that up to the fuses are raw. That is to say unshielded. These terminals are also right in front of one of the two drives, so perhaps thats where it lays.
One thing I have noticed is on the Diag. page, when I first click on it just after I have started Mach, my input pins all light up. They slowly go out , but not all of them. This is with no inputs even connected. I did have an estop setup, but I disconnected that while I'm sorting this out.
I am wondering if this might be the charge pump? Also, I can't actually disable the Estop switch in the config. I just removed the pin and port numbers, but the green check mark stays on. I find that odd. Possible bug?
I tried to capture the diag page, but things go a little crazy when I try that.
Enhanced pulsing is enabled. I think I have tried both ways.
This is version R2.0.065
-
I upgraded to the .069 version and all my leds on the diag page go out or they don't even come on. Even with both ports enabled. Port 1 state doesn't even light up. It still jogs the motors okay, so not sure what is up there. If I is a clue, I don't know what it means. ;)
-
Now I have something new that it hasn't done before. Both the X and Y motors turn poorly when I hit an X jog arrow key.
Sound noisey?
I narrowed this down to the el cheapo silicon keyboard.
I used the jog screen and that didn't happen, but I still get the bidirectinal movement.
I think I just need to break it all out of this case and setup with new wires on a table or bench.
I am not really satisfied with the banana terminals for the voltages through the case. They function okay, but not ideal. They are exposed inside.
What are some of the other type of through panel connections that are used?
-
I haven't had time to tear this apart yet, but I did try something else.
I don't think I have tried this before, but anyway, here goes.
In Mach, I swapped the pin numbers. When I run any of the three drives on pin 6 and 7, which was the good operating Z, they all run good. Thats not even changing any wiring. That is just swapping pin numbers in Mach.
I even wired one drive into the A axis and tried that with the same poor result.
Is this sounding like a port problem now?
-
I had exactly the same problems over the weekend. All axes drove well using pins 6 and 7 (Z) on my BOB. Y (4 and 5) could be tuned to run slowly, but X (2 and 3) was rubbish. As I had already tested the 3 steppers and drivers, my thoughts were that only the BOB or Mach3 could be the problem, so I substituted a direct connection BOB and all drives were OK. Even managed to draw Roadrunner!
When I find out the source of the problem and I suspect it is noise from the drivers or the BOB, I will post it in this thread.
Ian
-
Thanks Ian.
I tried this out again and my results were different this time or maybe I was simply not understanding what was happening the first time.
My Z runs good regardless of the pins, but the other two go both ways with one arrow key. Push the key several times and it will switch directions. Sometimes it does it with just two presses of the key.
I do think I just need to break this all out on a table and start from scratch to confirm that I'm not getting ground loops or electrical noise.
I do look forward to hearing how you fair with yours.
Thanks.
-
Problem located!!!!! I would holler YIPEEEE, but I don't feel good about not finding the issue sooner. Pretty much a no brainer.
I had took such care with this build. I used good techniques and went very slow checking everything multiple times.
On my first build, I had inadvertantly zapped a Xylotex when a motor wire touched the case. The only wire I didn't have tape on. It was fixed at no cost, but I really wanted to avoid that sinking feeling with this one.
Well....to the point.
I bought these connections off Ebay because I wanted an easy way to plug the motors in. They are nice and easy, but somehow two of the pins got bent to the side from the motors. Of course it couldn't have been the unused pin or a shield pin. They are on the other side of the plug. It was the same pin in both cases. One of the coils wire.
Anyway, problem solved. I thank you all for the help.
Now we know that if the motors turn both ways with just one key, it could be a loose wire connection.
;D
(http://leestyron.com/cnc/badplug.jpg)
-
Glad you are sorted and thanks for posting the problem, it may help someone else later on.
Hood
-
You wrote :-
"Now we know that if the motors turn both ways with just one key, it could be a loose wire connection."
That's one to note for future debugging. Great to hear that you got it sorted. It's the simple ones that get you every time.
My problem is definitetly noise. I can see it on the 'scope. A new BOB is on the way, I will post again if this cures it.
Ian
-
Thanks Guys. It was definitely an adventure. ;) I did learn a lot from it. I also shielded all my wiring. Thats good to have. Looks neat and won't hurt anything and may actually help prevent problems in the future.
The shielding came from Mcmaster Carr and may be worth looking into. It's very reasonable. It does expand larger than they list, but then relaxes to the size listed.
I have at least one more machine going after this one. That will give me three. The next one is a mill. I think 203V's and possibly a G100 will be on the menu for that one. I may want more than just three axis.
Thanks again and good luck with the noise.
-
I just thought of a test that would have prevented this.
Use an ohm meter to test the motor coils just before hooking them into the drives. I tested them on just the motors. Then I tested them with all the cables. I also tested the wires to the plug. Everything was fine. I neglected to test them once the cables were plugged in though.
I'm bloomin lucky I didn't blow a couple of gecko's.
I do feel much better now. Hopefully my ranting on here will help someone prevent this mistake in the future.
I knew it was something simple. ;D
-
Just an update on my noise problem. It was due to an incompatibility problem between the CNC4PC C11 BOB and the stepper drivers from Routout (UK). Arturo Duncan sent me a replacement C11 board but no change. He then sent a C11G, which is not fully optoisolated but has the spindle speed controller buit in. This cured the problem and my little mill is now running fine. I will post some photos in a new thread.
Ian Ralston