TC22 with Z axis vibration

[arsenix]

I bought a well loved TC22 last year and have been working my way through it. Beyond just scrubbing I went through the pneumatic, lube, electrics and fixed quite a few seals, leaks etc. Replaced pretty much everything rubber. Machine has been running quite nicely. The only minor issue I have left is that the Z axis seems to have a "rumble" to it. You can feel it in the cabinet. A vibration that is coming from the Z. I can see if it put an indicator on the head indicating the table. About a .001" shake of the needle (very fast). The posiiton indicated in the control doesn't vary. The axis position is completely stable and the movement is smooth. When jogging down the vibration goes away, and it gets a bit worse when jogging up. It varies a bit depending on position but is always there. I haven't noticed any surface finish issues from this, but I assume it will have an effect. X/Y don't have this issue, although I suppose the Z has a higher load when holding up the head. Machine is ~1998 vintage and has the larger Parvex motor.

I've been posting a bit about it on the PM forum but not a lot of interest over there: https://www.cnczone.com/v...8/#post3891574

Seems there are more TC22 users hanging out here. I wondered if anyone had any ideas on this. I haven't dug too deep into this one yet.


James

Similar Threads:

• Need Help!- 1996 TC22 with dx32 controller C-axis problem
• VM 16 X axis vibration
• cause of Z axis vibration
• X axis vibration

[bptcontrols]

James

Most likey to be Z-axis thrust bearing. At top of screw under axis brake. Do not remove brake without blocking head.

Parvex motors have a know issue: magnets come loose inside. if you jog up and down you may hear the magnets grinding and clattering. need quiet and listen next to column on side where there is a hole in casting, you can see motor.

Bart

Bart

[arsenix]

Ah ok. I see it in the diagram. Is there any good way to check the thrust bearing without disassembly? Is the PN the same as the X/Y thrust bearings? Oddly enough the diagram doesn't seem to have the PN for the Z thrust bearing. Just says "INA ZKLF". The X/Y are INA ZKLF 2575-2RS. These bearings are NOT CHEAP!

I'll see if I can hear magnets clattering. I presume that is not repairable?

Thank you Bart!


James

[bptcontrols]

James,
No way to check. I have a bad bearing as a paper weight on my desk. Looks and feels fine until installed and preloaded. You can take out and examine; one bad sign is if the inner race comes out its NG.

where are you looking I have a source thru a european mfg.

No the motor is not reliably repairable once magnets come loose from case, they are bonded on. Motor overheated and bonding fails. Magnets also weakened from heat stress. And when one comes off as soon as motor rotates it grinds it down, magnet NG after 1 rev.

No one rebuilds to factory specs, cant get parts. Some repair houses try to salvage parts to rebuild, success rate less than 50%.
There is an alternate US mfg that makes a custom direct replacement motor. I have been trying to contact to order a couple.
last time I saw new OEM Parvex, motor only, for TQ-22 over $3000. And thats if they will take the order. Only want to ship in quantity from EU and say 3 to 6 months. Still need encoder and connector. Delivered for install over $3500.

Parvex class motor only one capable of moving head, which is not counter balanced. Column casting opening for motor was sized for SEM motors like on X & Y. Not powerful enough; burnout after few hundred hours of operation. First 120 TQ22 had the smaller SEM motor. Lots of failures. Could not change casting. Found a small frame motor with several times the power: Parvex. Bridgeport replaced SEM Z-axis under warranty for the first 120 machines. It was expensive.

Bart

[arsenix]

Hm. I will examine it a bit more closely. If I lift on the head casting with an indicator on the screw will I see bearing play? Head weighs a ton no surprise this bearing would go out!

I poked around online for them. Prices all over the map. There are some for sale on ebay, NOS in USA for $300/ea. There was a guy who rebuilt his who sourced a full set (X/Y/Z) from a Chinese supplier. They claimed these were legit, ~$70/ea plus shipping. Seems dicey. He had no complaints but I didn't see a long term retrospective on how that turned out.

Hopefully my motor is still ok. That sounds like a road I would prefer to avoid!

[jdub63]

James,

I also purchased the new bearing from the Chinese supplier, they come marked "Made in Germany" ??. I haven't gotten mine installed yet, so I haven't tested them yet, but Ash is further along and had his running before he started retrofitting the control system; maybe, he'll weigh in. Let me know if you need a Z motor, I have a Parvex I pulled off and I'll be installing a new AC servo in it's place.

Jim

[arsenix]

Hm. If they were really made in Germany it is pretty suspect why they would be available for cheap in China, but it isn't impossible. A lot of machine tools get manufactured there, and they may get purchased in bulk at discounted prices to be competitive. Could also just be a straight up lie. Unfortunately pretty hard to know. There are a LOT of counterfeit bearings out there.

Are you doing a full retrofit with AC servos or installing new servo drives with the DX32 control?

[jdub63]

I'm doing a full retrofit, new servo motors/drives and Centroid Oak / CNC12 control system.

[Ivan_The_Red]

hey guys, i have a tc22 that im retrofitting and have a bunch of controller parts to sell if youre intetested, spindle motor, gearbox too.

[arsenix]

I made a program to move the Z axis up and down and listened next to the Parvex motor. No odd noises. No odd noises coming from the thrust bearing either. When stationary the belt is vibrating. I can feel it. The vibration goes away when I apply hand torque to the motor in one direction (my guess is less load?) and gets worse when I torque the other (more load?).

Since the encoder is on the motor I'm not sure this is likely to be the bearing? I suppose I could try removing the belt and jogging it around unloaded?

[bptcontrols]

Arsenix,
Vibration may be due to improper axis drive tuning parameters. Have you changed any axis tuning parameters: Axis Gain, Velocity Gain, Following Error Gain, Differential Following Error Gain, Direction Reversal Timer, one or 2 others. Parameters should be checked against original install and reset to factory.

Different axis parameters depending on version of software to improve performance. cant really mix and match. I did the tuning / testing when worked for Bridgeport. Spent months on production floor testing software and servo parameters to determine each ones affect on motion and its dynaminc range; find the sweet spot. When done machines could be tested on production floor by running test cycles and looking at position error, shown in Maintenance Page software, to see if something was wrong. I still have my original engineering note books documenting all my research. Factory never released any guidance or info on tuning. Just reinstall software or install update. Actually 6.87 or greater version of software have very good 'tunability'; each parameter affects a specific aspect of servo motion: sitting still, accel, decel, running at speed. If parameters are factory original and machine not run well there is another problem: servo DC power supply (approx 125vdc solid no ripple) the big capacitors are starting to age out and go bad, motor full of carbon brush dust or bad brush, encoder weak signal or noisey, poor grounding of motor, motor power or encoder cable shielding broken or not tied to ground in cabinet.

What version of software are you running; on startup see upper left corner of screen. Any version of software before 6.87 is what some early machines were shipped from factory. They were supposed to be updated upon install at customer site by dealer, many were not. Final release was 6.89. Much inproved, high speed machining, block look ahead, motion smoothing, s-curve accel & decel.

There is a very special configuration / set of tuning parameter I came up with for my machine, I call 'High Performance Option': faster and smoother motion block to block, more power at motor shaft, etc. Only works with version 6.89 and the high speed machining option, s-curve accel/decel, 50 block look ahead.

Bart

[bptcontrols]

Do you have any circuit boards. Controller, I/O, Tool changer control.

Why you retrofit.

Yaskawa spindle drive and motor is pretty basic in controlling +/10 volts, and spindle enable and fault.

What are u puttin on.

Bart

[bptcontrols]

Jdub63

Why are you retrofitting. I have done 3 Centroid Oak retrofits onto Bridgeport VMCs. Hate doing it but if thats wht someone wants and they are willing to pay its a done deal. Oak will work with Yaskawa servo and spindle drives. Wht are you doing for Centroid PLC tool changer programming.



Bart

[bptcontrols]

Made up this interface module so dont have to re-terminate original wires to OAK. Installed where AUF mounts, all the original green plugs go into sockets, Pheonix, on rail. Wires from center go to OAK.

Also power distribution for 24vdc to all relays and prox switches.

Motor mods to connectors so can keep original motor / encoder cables.


Bart

[bptcontrols]

Regarding axis thrust bearings.

Recently put 2 complete sets in Bridgeport VMC 800 and 1000. Owner found best price in Europe. Not sure of quality so ordered each set from different suppliers. 1st machine worked out good all the way around. But 2nd had major problem on one axis, excess backlash of 0.002, should be more like 0.0002. After reseat and reset preload could not get in spec. Checked ball-nut and screw all ok. Could not get rid of couple thousands end play. Took out bearings. Found one out of spec: inner race thickness, face to face, 0.001 inch thinner than outter race. All other bearings were same dimension face to face on inner and outter race.

After replacing bearings:
1) rechecked backlash and set in parameter table.
2) did axis lead screw calibration and set in parameter table.
3) Adjusted Siemens drive axis drive tuning.
4) Reset Bridgeport tuning parameters for In-Motion position following error of less than 0.001 at 1000 ipm.
5) Ran test part 20 inch bar drill multiple holes unequal spacing over 16 inches.
6) Checked on CMM. Every hole within +/-0.0001 desired position.
7) Ran rigid-tap calibration cycle and reset parameters. Tested with multiple 4-40 in aluminum. Perfect.

Dont get any better than this:
Bridgeport VMC 800 and 1000 run everyday; well cared for; as good as the day it was delivered 25 years ago. Cant get this out of a HAAS.

[arsenix]

Thanks for all the advice Bart.

I haven't changed any of the tuning parameters but I have only owned this machine for about a year and it had a long life with its original owner before that. SW version is 6.89.57, which was .apparently updated in 2002 (I have a floppy). This machine is from 1998 (serial #615). I have no idea if these are the original parameters. It is conceivable they were tampered with. I can check this floppy and see if there is a parameter backup on it. I can take a look at them.

I took the belt off the Z axis and just applied torque to the motor while the machine was running with my hand. It is silent when unloaded, vibrates when I apply torque in either direction. This seems like some kind of noise issue to me. I can check the supplies with a scope. I don't see the same problem on the other axes though (as far as I know... the other axes also don't hold up a massive heavy head all the time). I suppose I can pull and check the motor brushes pretty easily as well?

[bptcontrols]

6.89 software version is final factory release. This is good.

First thing to check is drive tuning parameters. Go into Maintence Page; bottom of start-up screen F5-MaintPg. Start by checking the parameters in the 60's X, Y, Z 'Velocity Gain' and 'Axis Gain'.

X, Y, Z Axis Gain affects how strong the system holds position or fights you to return to position when not moving.

X, Y, Z Velocity Gain affects position error when in motion.

There are several others

Around parameter #110
X, Y, Z Following Error (FE) Gain. how hard to push to follow move command.
X, Y, Z Differential Following Error Gain. Pushes harder with increasing position error.

Also several which affect X,Y, Z as a group: Hi FE Gain, Direction Reversal Time, and Hi FE Limit, Integral FE Gain.

Cant just change. Need to do in a specific order. If too high motion is unstable or low causes positioning errors, overshoot, or undershoot.

Will have to look up default settings.

Dont 'Install' the software from the floppy disk it will reset all the tool changer settings: height, orient position; and a few others specific to the machine.

You can do an 'Update' that only restores axis tuning and other standard settings for TQ-22.

[mactec54]

Hm. If they were really made in Germany it is pretty suspect why they would be available for cheap in China, but it isn't impossible. A lot of machine tools get manufactured there, and they may get purchased in bulk at discounted prices to be competitive. Could also just be a straight up lie. Unfortunately pretty hard to know. There are a LOT of counterfeit bearings out there.

Are you doing a full retrofit with AC servos or installing new servo drives with the DX32 control?

Most German Bearings there manufacturing is in China

[arsenix]

I had a chance to poke around with this a bit more. I removed the Z axis drive and inspected, checked the filter caps. They all tested correctly. I don't have a good way to test the big filter cap but it held 30v for ten minutes after i charged it up.

I put my scope on the power lines to the Z axis. I read about 20V Peak-peak when the axis is holding up the head. About 40v when I apply more torque with my hand, and it goes to near zero if I apply counter torque to hold the head up for the motor. I also confirmed I get similar noise on the X axis. I removed the belt housing and applied torque to the belt pulley with my hand.

This kinda seems like just an electrical noise issue on the DC bus. Maybe I should replace or upgrade the big DC filter cap? It has a 15mF, 250v cap there. The manual says the Parvex equipped machines got upgraded to 12mF 200V cap, so maybe this machine already had it replaced at some point? I am running this machine on single phase, so the power input may not be as smooth as if it was running on three phase? Although the DC bus only pulls in two phases anyway so it really shouldn't matter much. I could also try adding some extra capacitors right at the drive input and see if that does anything.

Maybe I should take a look at the PWM control signal to the drive and make sure it is applying an even signal.

[bptcontrols]

Measuring the voltage to the motor will not show anything of value. It is PWM modulated; width of the pulse is representative of current to motor which depends on load. Also there is 'Hi-Side' and 'Low-Side'. A transistor can fail, if so others in parallel are overload, overheat and fail.

Main DC bus should be a steady125vdc +/- 5 vdc. Will sag slightly under accel and pump up during decel. If there is a waver in DC bus or sags too much on accel main DC power pupply cap is failing.

The PWM drive board on the BMDC rarely goes bad. Same for drive card. They just flatout fail.

Ideal is 3 phase 20 amps mostly for spindle drive. Axis drive needs 2 phase and will pull 20 to 30 amps depending on load and speed. If you do full travel rapid up and down and DC bus voltage drops more than 10 vdc not getting enough current. Thus will show up as excess following error. Go to Maintenance Page and watch Position Error when rapid up & down. Must stay less than 0.010" during rapid else controller will slow down the motor.

If DC buss problem would also see it on X & Y. And usually not cause vibration.

An electrically noisey encoder, weak encoder signal, broken shield on cable from encoder. Enoder needs a good +5vdc. Measure at AXSBOB.

Check encoder on the Parvex motor; if Encoder Products model, silver can on end of motor, these were crappy and give all kinds of problems. Best is the Heidenhain, looks like a black can.

Also Motor armature shield must come as close to connection @ drive card as possible and tie to gnd buss.

The caps on the drive cards are special high performance: hi-voltage, hi-energy density, small package. Never seen these go bad.

You can swap the x or Y drive card to the Z-axis. wont be able to go full rapid but can home, jog, run at slow speeds.