FANUC 6MB servo error TGLS

1981 Kitamura Mycenter 2
FANUC 6MB
FANUC DC Model 5 servos

I was having some noise in the Y-axis and assumed it was a thrust bearing issue.
Replaced them and had the same noise.

Pulled the servo back off and ran it unattached and it seemed like alot of brush noise.

Pulled the brushes, blew out alot of carbon dust (I do this every 6-12 mos. and have seen more carbon dust the last several times)

Put new brushes in and was much quieter spinning by hand.

Reassemble, power up and now get SERVO 401 and the TGLS (Tach Gen. Lost Signal) led is lit on drive.

Google google google.... I pull the end cap off the servo to check pulse coder and I find ALOT of the carbon dust had made it into the end cap and settled right on the back side of the cannon plug for the pulse coder.
I blew all this out. Washed down with contact cleaner and blew it dry.
Checked inside the pulse coder and that was clean.

Reassemble...same error.

Disassemble.....pull the end bell off the servo and scrub all the carbon dust off inside. Polished the commutator by hand while in the motor. Scrape between the comm. bars until clean.

Reassemble....same error.

Swap the X and Y servos and the error follows the servo.

So suspected either the wiring from inside the end cap cannon plug to the coder or the coder itself to be faulty.

The cannon plug to coder board continuity checked out fine but that's with the wiring in entirely different positions from the assembled state.

I got a couple used pulse coders off internet (parted out from working machines...but who knows)

Spent over 4 hours resoldering cannon plug pins to completely replace the harness and pulse coder.

Reassemble....same error.

Could the servo motor itself be causing this error??

There are only a few servos on ebay that are 108 volt and 2000 pulse coder like mine...but I guess the next step is to roll the dice on one of them.

I just love replacing my 35 year old electrical parts with someone else's 35 year old electrical parts.

This machine still runs 1 or 2 jobs very well and pays its way....it's just these endless weekends of troubleshooting that gets it closer to the door.

Thanks for reading.

Connect your voltmeter to the motor power connection and spin the motor by hand. It should generate some voltage (how much depends on how fast it is spun). Positive when spun one direction and negative when spun the opposite. If you see no voltage you have a bad connection or a brush is not touching the commutator.

Vancbiker said:

Connect your voltmeter to the motor power connection and spin the motor by hand. It should generate some voltage (how much depends on how fast it is spun). Positive when spun one direction and negative when spun the opposite. If you see no voltage you have a bad connection or a brush is not touching the commutator.

Click to expand...

that's a good check, also check the tacho connections.

Thanks Vanc.....You nailed it.

No voltage could be read.

I pulled the brush caps and had to tug 3 of the 4 brushes out.
They were hanging up on the corner chamfers.

I must've shaved the 1st one when checking for fit after I cleaned the end bell.

Reassembled, power up...no error.

Vancbiker said:

Connect your voltmeter to the motor power connection and spin the motor by hand.

Click to expand...

High end work that. I congratulate you. I'd almost forgotten about spinning a DC motor and turning it into a generator. I do it with tachs and a CRO. But seldom on a D.C main Commutator and Brushes.

Regards Phil.

Bob E said:

Thanks Vanc.....You nailed it.

Click to expand...

When the 6 control is powered up it sends a tiny motion command to the servos and checks for feedback. If it does not see any feedback, it goes into e-stop and generates the TGLS alarm. Works even when there is a no tach due to some manipulation of the encoder signals to "simulate" a tach signal.

Vancbiker said:

When the 6 control is powered up it sends a tiny motion command to the servos and checks for feedback. If it does not see any feedback, it goes into e-stop and generates the TGLS alarm. Works even when there is a no tach due to some manipulation of the encoder signals to "simulate" a tach signal.

Click to expand...

Could this be temperature related somehow? We are getting this TGLS alarm at start up. If you power up the NC but leave the e-stop in for an hour or so, then release the stop, it'll start up. If you release the estop too soon, it'll get TGLS.
We are also having a condition that will shut the machine down in the middle of running, it acts and resets like an e-stop. I have not investigated to see if that has TGLS alarm in that condition, I'm thinking it doesn't because someone would have noticed a red light on a board by now and brought it to my attention.
Maybe they are related?

I'm currently going after cleaning the motor brushes.

A failing power supply is my initial thought. Often if troubles go away after turning a machine on and off multiple times or leaving partially powered on for a while bad capacitors in a power supply is the culprit.

Vancbiker, are these capacitors that you speak of in the control units that I have boxed in green? They would be behind the top board, the big ones correct?

The power supply for the control is at the bottom of the cabinet. Next to the red circle.
Has the cover with the lightning bolt on it.

SeikiCorp said:

The power supply for the control is at the bottom of the cabinet. Next to the red circle.
Has the cover with the lightning bolt on it.

Click to expand...

thanks! I just checked the big capacitor under that cover and it checks good. I also checked the three under the top board in that particular drive that I'm getting the alarm in, they are all good also.

They have switching supplies for the control portion.
They really have to be checked with a scope for output waveform. ( ripple )
It is the ESR that dies with old capacitors. Then they pop or just quit being good capacitors.
They may hold X volts for so many seconds which is fine for a linear supply, But at the operating frequency of a drive or switch supply their value might be a magnitude away from what they should be.

Pull out the supply, Put it in the freezer for 2 hours, Reinstall it, Test it quickly while it is cold.
Most of the connections on those are a plug-in on the outputs. Easy for a quick test.

Thank you seikiCorp, I have a scope, where should I hook up to run a test?

Casey25 said:

Thank you seikiCorp, I have a scope, where should I hook up to run a test?

Click to expand...

There are multiple outputs from a supply of that era.
The major ones are +24v, +5v
The 'weird' ones are generated on the control board. The -10v or -12v used for analog servos or spindle drives. That depends how the builder setup their control. That will show up as spindle not running the commanded speed.

Clip on the output wires. if it is a dual channel scope, input1= 24v, input2=5v

Do you have a close up photo for the power supply?

Casey25 said:

Vancbiker, are these capacitors that you speak of in the control units that I have boxed in green? They would be behind the top board, the big ones correct?

Click to expand...

SeikiCorp is pointing you in the right direction.

Here's a pic of the power supply, thanks for all your help!

It comes out as a big slab. 'Brain' board and power board.
Loosen the screws and slides up. This is the first time I have seen one without quick connections on the output side.

I have labeled the provided photo.

If you have access to one of those 'cold guns' or liquid C02 you can test it in the machine.
The fault can be in either the brain or the power board.

30+ years is a good lifetime for power electronics.

thanks for the help, I'm going to get my scope and check those output waves. Should I, or can I test that row of output capacitors?

Do you have a ESR tester? Those are several thousands of dollars.
A capacitor tester on your handheld meter is the equivalent of using a test light to diagnose a supercomputer.

You can spend so much time and effort to trace a intermittent issue.
It would be better to replace all of those output capacitors at the same time if you are so inclined. Watch the temperature ratings, voltages, and ESR of the new capacitors.

Do you have a part number for the assembly?
I have access to the Fanuc e-parts system. I can lookup if it is available.

I was trying to help you prove out it was the supply and not a drive.

I'm pretty new to the oscilloscope, I can't seem to get the wave settled down. It jumps back and fourth, and my peak to peak voltages seem like they are much higher than they should be. Hopefully you can view the video links below.
Do you think this could be related to my other problem of the machine shutting down in the middle of running a part? No alarms or anything. See my other post, maybe we should move there for this discussion.

24and5V
5V
24V