Cause? Runaway Servo when inch move requested

[sparkgirl]

Problem= Z axis does rapid drop (or rise) when requesting 1.0" or 0.1" move. Direction of move is correct. This occurs whether using WinCnc control screen "button" or a G-code inch move request typed manually into WinCnc. I have had a two year hiatus from my table, therefore all the steps to rule out causes are still hazy.

My current setup is--Dual table set up using PracticalCNC built (6/2006) servo controller box for XYZ. The Z axis is ball screw with round rails. I do not know how to describe the axis driver boards/cards, except to mention that (unfortunately) they are wired in a series with jumpers so only Z is directly wired. 3 MCG (DC brush-I think) Servos, Nema34. Uses Heds Rotary disk style encoders with (unfortunately, again) a 10-pin extension PCNC soldered on for RJ45 jack. Runs on Windows XP home on non-networked computer. Was a dedicated computer, now has lots of extras I need to uninstall due to temporary home use. Router (no spindle) not started in WinCnc, just torch is, (that's if I could attach router without it dropping into spoil board). WinCnc is v2.2, not upgraded to 2.4 yet. CPU has PCI card with attached daughterboard installed for motion control. Daughterboard was new and tested (supposedly) 6/08, just before table was stored. CPU tower case is clean. Servos were left attached to gantry during storage--experienced humidity and freezing temps for sure.

This is what I have tried/tested so far--

• 0.01" & 0.001" moves okay with software "button" execution.
• 1.00" request will move about 4" in Z--X Servo has runaway when X cables in Z.
• Swapped encoder readers(RJ45)no change (?rules out encoder)
• Visually inspected mylar disks, looks clean, no warp.
• Didn't remove or swap rotary disks.
• Performed tests with "dual" controller box set to joystick-manual torch-no Z, software mode for router, and ATHC modes-same response with Z.
• Did test runs with a new cat5 cable too.
• Swapped the X & Z cat5 encoder and power cables at servos and box drives-
• Z worked fine when run from X-driver board.
• X had major runaway, more than Z's 4" (?higher feed rates) when run off Z-drive. Had to abort before I bent the skinny wimpy X shaft
• Didn't switch the phoenix-end jumpers out of their series in drive boards during X & Z swap...hmm, reminds me servos are not grounded, will do that next.--okay done.

Trying to recall all the program codes, +equipment glitches, and too little sleep have rendered my brain temporarily immobilized, running in only 2D, just like my table. Would someone please assist as an auxiliary brain? It's likely I need someone to point out the obvious clues to me...other than the need for a nap.

[sparkgirl]

Got some sleep and have thought to add--
I have also gone over WinCnc's

• cncscrn.ini & wincnc.ini file to check that input lines hadn't changed since last worked correctly.

Thanks in advance for any input on what to try next!!

[Mariss Freimanis]

Whose drives are you using? If they are ours (Geckodrive G320s, 2.5" by 2.5", black anodized, red colored lettering), check to see if the encoder cable has become disconnected. That is a sure-fire way to get a servo runaway.

Mariss

[sparkgirl]

Thank you for your reply Mariss, I have seen good ratings for Geckodrives and your support on the forums. Evidently, my table was equipped after Joe's "supply" of Geckos ran out. The 3 drives installed in my controller appear to have the black anodized shield on only one side, not encased. I view no labeling except #s (2005) 2010-0504 on green board. They appear to have only an 8-pin connection.

...ooh, just found 'em on the web, they are Rutex R2k ServoDrives-R2010 for brush servo motors. When powered, I see a red light on the Z-drive, and green lights on the X and Y drives. Should this be a clue for me?

I just went over my grounding wire set-up and moved the earth ground to a new location--unused screw in the base of controller box, scraping away any paint first. My husband thought he remembered something similar happening in the past prior to any separate grounding. He took a quick look, (possessing more electrical know-how) and gave my grounding the okay before leaving on his fishing trip---still no change.

The only new point to add is that if I slow the feed rate way down in WinCnc, I can reduce the servo from total runaway to just jogging much farther than the distance parameter I input. I have not been running actual files yet, just doing small jogs, transit moves, or single line entries. I remain overly cautious.

I think this means the drive is getting some signal from the encoder?? I remember the time my hubby unintentionally bypassed encoder and slammed the drive so hard into an endstop that the grease zerk for bearing car broke off. (I was not in the shop, tho he described it well!)

[Mariss Freimanis]

sparkgirl,

There is a long and sad history when it comes to Practical CNC and us. Joe Rathbun of Practical CNC fame burned us for $6.5K in drives. He was (and is) an out and out crook.

Rutex was a good competitor of ours but they went belly-up for technical reasons. Once Joe couldn't get drives from us he switched to Rutex who promised more than they could deliver.

All that is inside baseball and you shouldn't care about it.

My guess is if I was in your position would be to focus on the drive with the red LED. Try swapping it with a drive that works. The Z-drive showing a red LED is problematic. Joe wasn't a technical guy and he never understood the importance of motor gearing.

The Z-axis is the one that holds a weight against gravity and it's badly stressed if the motor gearing isn't right. Joe never worried about such things so his machine design reflected it. You have to deal with it.

Usually a red LED indicates some problem. It is a convention we use and I assume Rutex and others used it as well. Download the Rutex user's manual and find out what it means.

I'm not familiar with Rutex drives. My job is to design Geckodrive motor drives. The only help I have to offer is general. Anytime a drive behaves OK at low speeds but acts crazy at high speeds is noise related. Higher speeds involve higher currents, higher currents generate more noise than lower currents. If the grounding from motor and encoder is incorrect, you will have problems.

Usually these problems manifest themselves as incorrect positioning. To have a servo runaway means the problem is very severe and therefore easy to find and correct. Look for loose grounding wires from the motor and encoders.

I winced when you said the Rutex drives were daisy chained. Joe wasn't technical and he never believed the good advice he was given. He chose to save 2" in wire rather than wiring his machines correctly. Besides being a crook, he was cheap. Cheap is worse.:-)

Once you figure out you current problem, you might want to consider rewiring the machine correctly. Correct wiring means 3 things: "Star" power distribution, single-point ground and ground return. Let me take them one at a time:

1) Star power distribution. Just like sun rays from the sun. If it requires power (drives) each gets their own "+" and "-" wires from the power supply. Each drive gets two wires from the power supply to the drive. Simple.

2) Single-point ground. A single wire goes from the power supply to the chassis of the control cabinet. That is the only connection to the control cabinet. The control cabinet doesn't provide any connection to anything else anywhere. Ever.

3) Ground return. Every motor must have a wire coming from the motor case going to the single ground in (2). Two wires to the motor armature, a third wire from the motor body to (2). Do that for every motor.

Encoder cable ground. Terminate the encoder cable shield wire at the control box end. return the shield wire to (2). Do not connect the far end to anything! Leave that end unconnected. It's important.

You will wind up with a very busy connection to (2). Use a #6 or #8 screw to the chassis. Use as many #6 or #8 solder lugs (the things with the #6 or 8 thru holes at one end and wire solder holes at the other end) as you need.

You will get for your trouble a control cabinet and drives that are absolutely free of noise and trouble-free.

You have a mess on your hands right now and it's not your fault. They aren't our drives either but this is how you correct it and make it better than it was before. It's how you take it from flaky to perfect.

Mariss

[sparkgirl]

Thank you Mariss, for taking the time to explain the best method to correct both my immediate, and my larger overall problem, as well as doing your best to "soften" the blow. I applaud you for going far "above and beyond" the call of duty!! That is the great thing about this forum, people sharing the wealth--of knowledge, that is!

I happen to have relocated quite close to MI Tech, a university with a highly credited electrical engineering program. I will put a request out there for some assistance. Although you explained my needs very well, I think I will save more time and headaches with hands on tech experience to execute your suggestions. I have stretched my pea-brain immensely so far as a pre-owners guide/manual PCNC owner. Never dreamt I would learn so much background stuff to run this table...I was led to believe it was a plug-n-play deal, ha-ha! I've already gone astray with my daisy-chained grounding, using eyelets to connect at one mounting screw of each servo to the next.

I will swap out a drive next, I held back from this so far since the problem Z drive is the one with the wire connections, and I was timid with having to cut and reconnect. I have ordered a slew of back-up parts from another helpful member on this forum, and they should arrive in a few days. I can not yet afford to switch over completely to a new system as the seller of my parts has done. Wish me luck. Gotta go waitress for 10 hours now to help pay for parts--Yuk!