X motors reversing direction at random

[lto]

I'll post here much of what I emailed to shopbot in hopes that someone has a clue about what is going on.

You may recall that back on Saturday February 26th you helped me via phone while I was switching control boards. We were hoping another board might alleviate some of the chatter I was seeing in my edge finish. On a side note I have since improved the edge a bit with machining and holding strategies and also very low move speed cleanup passes (.5 ips).



Upon having installed the control board on 2/26 2005 we had the strange occurrence of the X motors moving in the reverse direction of that shown on the screen. I then tightened the screws on the x driver plug and by all appearances this seemed to be the solution as we didn’t see this happen again for a few weeks. I think it was two weeks ago (I know it was on a Saturday) that this happened once again. I thought it might possibly have to do with static from the dust collector as we were cutting plastic at the time. We added some more grounding to the dust system and had no more problems for the past two weeks. A side benefit was that the collector barrel ceased giving wake-up calls whenever brushed against.



This brings us to today. We were starting to cut a previously cut file when the machine lost position. We repeated our attempt six times and noticed that only the X is affected. This reminded me of our experience on the past Saturdays as noted above. Sometimes the carriage would seriously shudder as though it were shivering from cold. By repeated observation I came to the conclusion that this was because of rapid forward/reverse change of the X axis. I checked connections in the control box, everything appeared fine so I just wiggled them, then tried the file again, without success. It doesn’t seem logical that the machine has run without this problem for two periods of several weeks each. Also I question whether the control board connection to the X driver which I had formerly suspected is the culprit as both X motors are reversing in unison, (no racking that I could notice).



Another note; I have installed the new software update and like the added features. So far I haven’t noticed a difference in smoothness of diagonal movement



Hopefully this reversing of the motors is not unprecedented, I can do without the Shopbot for today but I certainly need it back up early next week.



Sorry about the long-windedness, but I thought recalling some background information might help in diagnosing.

Thanks to everyone who has read this. At first glance it might be determined that the Shopbot is averse to working on Saturdays. However it has behaved fine on a few other Saturdays and I'm sure there is a more logical answere. By the way this is a three month old Alpha tool. I'll be glad to hear any and all suggestions.

Thanks LTO

[Brady Watson]

LTO,
Are you running a PRT or Alpha?

What software do you have installed on the PC? Is power management running (turn it off), Quicktime or anything else that could take focus away from SB3? I intially had some of the same issues and it was cleared up by turning power management off completely ~ using an 'Always On' profile for everything.

Other than that, (I am guessing that you have an Alpha)...if you said that symptoms went away after tightening the X-motor plug(s), then this would lead me to believe that there is arcing or a bad connection at that plug. The Alpha motors are supplied 170vdc, compared to a PRT @ 40v, and the possibility of arcing is much higher on an Alpha. It may be that a loose connection is arcing and sending feedback to the driver, wigging it out and making it reverse direction from time to time.

Just a few thoughts...and places where I would look.

-Brady

[fleinbach]

Louis

I forget, are you running a Spindle? If so is it the 3 or 5HP one?
I too had some simular problems. They where cured by grounding both ends of the spindle motor cable.

[lto]

Thanks Brady and Frank,

I am running an Alpha with a 3HP spindle. I will check out everything that you have suggested. The thing that baffles me is that the machine ran without any of this for the last two weeks, now it has reversed direction seven times today. Once at the very beginning of the file, the other times as much as five minutes into the file.This photo shows where coming around a curve the machine continued on the correct arc but in he opposite direction. Then it apparently sensed it was off and tried correcting things again just before it was shut off.


3840

I have already been contacted by Grant at Shopbot. I have deleted the ini file per his recomendation and started over, similar results. Shopbot will get back to me Monday to hopefully tame this animal.

[fleinbach]

Louis,

You have the same spindle as me. I was having the same problems you are having intermitantly until last month. At that time for some reason one file I was running failed consistantly. Ted at Shopbot asked me to check and see if the Sheiled wire was connected on both ends. He said they usualy send the spindle with only one end connected, which works for most people but for others it works better if both ends are connected. Well I found the sheild was not connected at the spindle and I reconnected it. I have not had one problem since then.

[Brady Watson]

Yep...now that you mention it Frank...shielded cable can be a bugger on these machines. Generally speaking in regards to CNCs, there needs to be a common grounding point for ALL grounds on the machine. Shielded cable is USUALLY not connected on both ends because this creates a ground loop that causes all kinds of gremlins. It is the role of the shield to act as an antenna and carry any leaking energy (static or otherwise) to the common ground. Since it is only connected on the one end, it naturally follows the path of least resistance and goes to the common ground. If it is connected on both ends...then the path of least resistance could be anywhere BUT the common ground. Since the problems went away when you reattached the sheild on the spindle end, this leads me to believe that the spindle and Z axis are not grounded as well as they should be. What about a woven ground strap bolted between the 2?

-Brady

[gerald_d]

When I saw another recent model SB this week, I was surprised to see that V-rollers are still expected to form the ground path from the z-axis, through another 2 sets of rollers, to the common ground.

[lto]

Update on X axis reversing problem;

I received another email from Grant at shopbot around midnight Saturday night telling me that they would do their best in getting me up and running Monday, He sure is on the job. I didn't get to try the machine until this afternoon again. It did the unexpected reverse thing during the homing routine without the spindle running. I called Shopbot after 1PM and tech support was now reading my Saturday email! Sorta makes one wonder if they're understaffed or just have too many fires to put out? Anyway I was called again and recommended to reinstall the firmware. I did this and cut for 1 1/2 hour plus without a problem. Then just as I was getting comfortable and no longer standing over the shift key, it reversed and lost position again, cutting a new line through the work piece and into the holding jig. By this time its to late in the day for Shopbot to ship another control board, (which we don't even know will solve the problem).

Frank L., before starting the machine today I checked on the spindle wire grounding. The sheild was grounded at the variable frequency drive only. I would have grounded it at the spindle end but there was no shielding visible at that end of the cable. Before cutting the cable back to get to sheilding I ran this past Frank @ Shopbot and he recomended against it. The ground wire is connected to the sheild at the VFD end and connected to the ground bolt at the spindle end, so in this way I suppose the shield would be grounded to the spindle. Since I did have the malfunction happen even without the spindle running it does seem logical to outrule this one as a cause. Just curious, your interrmittant loss of position problem must not have caused much lost material or you couldn't have put up with it for so long.

Again, thanks for the suggestions I hope to be able to soon post a solution.

[fleinbach]

Louis

I'm a little confused because you state that "there is no sheild visable at the spindle." (Just as mine was.) But later you state after talking to Frank that the "ground wire is connected to the ground bolt at the spindle end" If you are talking about an external ground, I wouldn't count on it. Grounding is very strange.

I slid the plastic nut back at the spindle and cut a slit about 3/4" long. this gave me enough room to pull out the sheild drain wire. I soldered a 4" piece of wire to it and pushed it back through the plastic nut inside and connected it to a ground.

You can't lose anything but a little time trying this. It apsolutly stopped my problems. My problems while cutting happened every week from last fall until it was fixed several weeks ago. It was a bit of a problem but not enough to stop me from recutting the bad part. Most of the time it would only ruin 1 out of maybe 2 or 3 dozen parts. All I had to do was re-cut the bad part. The cost wasn't that much it was all MDF. Also I was able to cut 2 or 3 files before I would get a failure, that is until the straw that broke the camels back. Suddenly it happened on attempt to cut one particular file. That's when I had to find the problem. But I am dead certain hooking up that ground fixed my problem because after I hooked up the ground I ran the file several times air cutting to test it. Not one failure. Then I removed the ground wire again and ran the file. It failed every time with the ground off.

[lto]

3841


3842


3843


3844
This last one is called Shopbot introduces $55 "Super O" bit to screw.

[lto]

Frank,

Thanks for the thorough explanation, I am certainly trying this.

[lto]

I just tried the grounding similar to what Frank explained. I wrapped shielding tape,(wire mesh)around the stripped end of the spindle cable. I grounded the other end to the spindle housing. Maybe I missed something here? The X started stuttering 30 seconds into the file. I give up for tonight.

[richards]

LTO,

Since you've already tried re-grounding the spindle, I assume that you've followed Gerald's advice to tie the X-Axis, Y-Axis and Z-Axis to a common ground. (I use the braided wire commonly sold to ground the plastic hoses of a dust collector.) In addition, make sure that the plastic hose on your dust collector and/or the plastic hose on your ShopVac (if you use one for vacuum hold-down) is laced with that same braided copper wire, and that ONE end of the wire is tied to the common ground point where the X, Y, and Z axises (axi?) are tied together.

If you've already tried that and still have the problem, I would start to suspect the controller card. (As usual, I'm away from my shop so I can't take a quick peek at the oriental stepper controllers.) All of the oriental motors stepper controllers that I've used, CSM and Vexta, use TTL logic input to the controllers to control the rotational direction of the motor. In a electrically noisey invironment (such as the electronics drawers that I formally built for the Kodak S-printer), I often had to add a 4.7k pull-up resistor to the TTL line feeding the input signals to the stepper controllers. At times, I even had to add a small capacitor to catch the short spikes. The resistor guaranteed that the input voltage was 5vdc instead of the more normal 2.2 to 3.5 volts dc normally found in TTL circuits. The 4.7k resistance was high enough that a normal TTL circuit had no trouble pulling the voltage to near 0 volts (specs at that time were 0 to 0.7 volts for the low level and 2.2 to VCC for the high level).

I'm not suggesting that you get out your soldering iron and resistor assortment to modify the controller board. What I am suggesting is that you might be one of the "lucky" ones that has lots of "EFI" gremlins playing the "gotcha" game with your electronics. If that's the case, you'll probably have to find a way to reduce the "EFI".

Electrical noise is normally VERY hard to fix simply because it is almost always random in nature. At times I've even resorted to adding temporarily a little D-type flip-flop circuit (7474 chip or equiv.) to an input line, with an LED tied to the output of the D circuit that would light up and stay lit when those little spikes occurred. Again, I'm not suggesting that you build other circuits to troubleshoot your existing circuits, but it's a tool that I've had to resort to using from time to time.

One last thought, a electrically noisey stepper controller could also self-generate spikes that could interfere with its own circuitry.

[lto]

Thanks everyone for your suggestions. I tried numerous things including grounding the wire sheild at the spindle end, (per your instructions Frank). The problem finally disappeared upon installing another control board that Shopbot sent me.


On a side note; I tried holding this blank on a mask 4 3/4 x 118" using a 1 HP vac pump. I was leaving a skin and the vacuum held at 20"Hg. However the pump,(mounted on a 20 gallon tank) needed to run continuously. Is this normal for a part of this size? I'm trying to determine what size pump I'd need to handle a dedicated jig allowing cutting completly through on parts like these. Would something like what Shopbot sells, (generating 14" Hg provide sufficient hold on parts with only 10" of area inside the gasket?
3845

[billp]

LTO,
I have been doing a LOT of research on vacuum over the last few months. I've tried rotary pumps, reciprocating pumps, Shopvacs, etc. As you'll see in another thread on this Forum I have even devised a small ( 1/4 hp) rig that I use with PVC "pods" to hold down large objects just about anywhere on the table.
In all of this testing I have found that the one factor which makes the MOST difference in performance is the use of a proper "gasket" material. I started out with the refigerator tape sold at the Depot, etc., even tried an old wet suit (for the neoprene...). The BEST results ( as I recently demonstrated at the Oklahoma Camp) were produced by a closed cell gasketing material made specifically for CNC usage. I have been using the tape and sheet products from a company called "All Star Adhesives" ( www.allstaradhesives.com (http://www.allstaradhesives.com) ) I first saw them at the IWF show in Atlanta last year and followed up on their products. In fact I have invited them to come to this year's Jamboree to act as both a vendor, and a presenter during my session on "Vacuum hold downs/jigs".
I use the gasket for cutting out lithophanes, and wooden parts primarily. I am able to generate 25" of mercury with my rotary pump rig, and by using the gasket tape ( 1/16"thick in this case) it takes over 5 hours to "leak" down to 18"of mercury.
I do NOT think it is necessary to use a large pump to generate sufficient holding power. In fact I think that with the Fein III vacuum sold by Shopbot , AND the gasket tape, you could generate more than enough suction to do the job you suggest. There is a picture of the rig I am currently using on my web page ( www.baycraftdesigns.com (http://www.baycraftdesigns.com) , under the Shopbot section). I think you should investigate some of the more efficient setups before going to the larger ones...

[nappy1]

You guys have got to start using PODS. Bill Palumbo can verify that.
I don't care who you buy them from, me or sombody else but you have got to try it.
I cut 14" Corian ovals with 1 8" x*' pod and a $59.00 pump.

[ron brown]

I've been cutting 7" maple circles into coffee jar lids using a $16 vacuum motor and weatherstripping for gasket. I've never measured the vacuum but I suspect it is near 100" water - that would be close to 7' or 8" of Mercury (Hg). I don't worry about leaks or too much about gaskets.

There is no magic in these setups. One does need to be aware of the pitfalls and problems that vary from inadaquate vacuum volume to porus material. if one need to know the vacuum on a part, it needs to be measured at the part, not the pump. This I got to visit yesterday when a fitting would allow the pump gage to show 15" Hg when in open air. I was vacuum impregnating epoxy into some mesquite burls. Once the chamber was sealed the vacuum went to 28" Hg. The orfice into the chamber was ~0.100". I was amazed I could pull any vacuum through an orfice that large with a 3/4 HP vacuum pump.

Ron

[gerald_d]

The gasket material has a second function in addition to sealing leaks - it creates friction (or grip) to stop parts from sliding around.

If you suck ice to ice with the biggest pump in the universe, it will still slip. But, if you add a bit of grit, the picture changes drastically. The friction properties of the gasketing may even be more important than the sealing properties in our applications.

The last couple of posts should probably be in another thread.......