[sortafast]

So I have a Tree 325J. It has been a solid machine but a bearing in the Z axis servo died a very miserable death. Like an idiot I jumped right in before googling stuff and tore the thing apart (very carefully I might add), cleaned it up a bit and replaced the offending bearing. Now it can actually turn with out squeaking. Anyway, I reassembled it and now I do not believe that it is getting proper feed back from the optical encoder, the resolver/tach thing, or both. When I fire up the control and get started on the start up procedure, once I clear the Shut Down E-stop faults, the servo just starts spinning then throws the code Z Axis runaway or something to that effect. Trying to figure out what the deal is here. I am thinking that maybe stuff is just somehow out of adjustment. There seems to be rather little on the interwebs that I have been able to find that has been of any usefulness to this situation. The Servo is an MSI corp MTE 404-500-MA that I do believe is a brushed DC servo. Upon taking it apart I noticed that there are several pieces to the body. Pop off the non-shaft side cap and there is an optical encoder with 500CPR incremental disc. That sits attached to a secondary cap. The screw holes for the mounting screws for this second half are slotted for what I am thinking is some sort of adjustment to align the optical encoder with the tach/resolver thing that is hidden under this second cap. The Tach/Resolver thing has 4 small brushes in a plastic housing. 2 of the brushes have wires going to them. These are 90° apart. then there are wires connecting those 2 brushes to the other 2 brushes ( am assuming to the brush 180° am too tired to check right now, but I think that makes the most sense). These contact something that looks similar to the copper part on this (photo pulled from google):

Attachment 311898

This Tach thing looks like it would be rather hard to screw up unless one of the wires isn't making contact (which one was loose and I took care of that). So the thing that I am left with, is that is the alignment of the optical encoder dependent on an adjustment relative to the tach? Like does it have to be in phase with it some how to make all this work together for the control to do its thing properly? I am assuming that this might be the case due to the slotted screw holes. Anyone have any insight into this? I would really hate to have to buy a new servo as it looks like it might be a bit of a pain. Would rather try to get this thing going again with what I have. Otherwise its time to start looking at either dropping some $$$ on a conversion or selling it off and getting something else. Thoughts?

[mike_Kilroy]

Z Axis runaway or something to that effect.
typical if you rewire and reverse ANY wires on armature, tach, or Encoder. Did you take pix before? go double check them against how it is wired now...

The Servo is an MSI corp MTE 404-500-MA that I do believe is a brushed DC servo.
You "believe?" Does the armature of the motor have BRUSHES? If so, then yes, it is DC brush motor. just googled it; looks like dc motor with brush caps all around...

optical encoder with 500CPR incremental disc.
as long as you did not mess with the disc set screw adjustment, you probably did not harm the encoder.

The Tach/Resolver thing has 4 small brushes in a plastic housing
we call that the BRA (brush ring assembly. nothing to adjust here.


This Tach thing looks like it would be rather hard to screw up unless one of the wires isn't making contact (which one was loose and I took care of that).
true unless manhandled and broke something. It will run without 1 brush, just might be more jerky, certainly not cause of run away.

So the thing that I am left with, is that is the alignment of the optical encoder dependent on an adjustment relative to the tach?
NOPE. no alignment required or either on dc motor. see my first comment above: either you broke a wire along the say and have a MISSING tach or encoder signal or you have reversed a wire - most likely.

You CAN send the motor to folks like us who repair them daily, but I think you would be wasting your money as it runs and 99% likely you just reversed a set of wires along the re-hook up part of the repair.

[Al_The_Man]

Agree it appears to be loss of tach runaway, this can occur either because of reverse polarity or no output, give the motor a spin and see if the tach outputs a DC voltage on a meter, the level of which is often on the motor plate, 20v/krpm for example.
If it outputs, try reversing the tach leads.
Al.

[mike_Kilroy]

Just replied to your PM; I think keeping it here is best so you get all the ideas, like Al's good one... that said, here is what I added;

HI Dave, I replied on your post instead. I will write more there too...

Sure we can check out your motor and repair it if bad, but 99% chance it is a wire issue - Al added what I did not write as most likely - unless you swapped the 2 motor armature leads - that too will cause the run away, as well as a missing wire on the encoder feedback (dont know your controller cnc - it MAY just shut off if missing encoder wire, but I cannot say - try checking tack like Al said with voltmeter, then double check your 2 armature leads, if none of those, then unplug the encoder from cnc and see if it does same run away or if it shuts down on enc fault. If tach, armature, checks prove good, and pulling enc connector gives same result, then you have a missing encoder signal - check those wires better....

if at the end of the day you find nothing and want to send us the motor for evaluation, ok. send info to Dennis at sales... - we do not charge for evaluation but you will be paying for shipping both ways.... we can then tell you A) the motor/tach/enc is all good, or B) something is broke and it will cost this much if you choose to have us repair it.

[Al_The_Man]

If you have one of the Tree mills that were fitted with Gould servo's, these often had 9.5v/krpm tachs on them.
One easier way to check the motor, is to remove it and run it on the bench with an automotive battery, check the tach for output both direction ±v out.
Al.

[sortafast]

optical encoder with 500CPR incremental disc.
as long as you did not mess with the disc set screw adjustment, you probably did not harm the encoder.

I took the disc off of the shaft as I had to in order to get to the offending bearing. I am hoping that I didn't mess this up. I found the data sheet for the encoder and it looks like they still make it, but I am not 100% sure if they still make 500CPR disc for it. Haven't looked too hard though. Looking at the data sheet there is some sort of adjustment/alignment tool to get it right. I may just order up an encoder and the tool and see what I can do there. I am hoping that it's not ruined, just maybe a bit out of alignment.

As for the Tach, there are only 2 wires and I didn't mess with those. Didn't take them off, but one of the terminals was a bit loose. I tightened that but I don't think that would have caused this as there was still decent contact there. I will bust out the DMM and see what I can see with that.

[Al_The_Man]

As long as you did not scratch the encoder disc, it should be possible to align it centre of the opto sensor.
The encoder is mounted independent of the encoder, with the tach usually going to the drive and the encoder to the control.
It is very tricky but I have used an ordinary meter to see the A. B pulses outputs although a 'scope is better.
If you get it working the only thing that may be out is your home zero if it uses a L.S. and encoder Z marker pulse method to home.
Some systems such as Fanuc have a grid shift parameter to adjust this in software.
Al.

[sortafast]

As long as you did not scratch the encoder disc, it should be possible to align it centre of the opto sensor.
The encoder is mounted independent of the encoder, with the tach usually going to the drive and the encoder to the control.
It is very tricky but I have used an ordinary meter to see the A. B pulses outputs although a 'scope is better.
If you get it working the only thing that may be out is your home zero if it uses a L.S. and encoder Z marker pulse method to home.
Some systems such as Fanuc have a grid shift parameter to adjust this in software.
Al.

Zero is not an issue with this machine. When it shuts down you lose the zero every time. It was the way it was designed. No home switches, just limit switches. So it's stupid simple. Kind of annoying that I have rezero my fixtures and such in the morning but its only 3-5 min to do so it's not terrible. You just reference each axis so that I finds the "ticks" on the encoder (usually less than 0.050" of movement) then that is your home location for the day. It has it's drawbacks, but so far it has been a great machine and pretty darn reliable.

[Al_The_Man]

Seems a really oddball method of zeroing, may as well use the power-up position?
As you may know the common method is to rapid to a home limit, then seek the one/rev encoder Z pulse in slow.
Al.

[sortafast]

Seems a really oddball method of zeroing, may as well use the power-up position?
Al.

Thats pretty much it. You make sure the table is aligned to a couple arrows and that the Z is all the way up but not hitting the limit switch. Then you reference each axis so that it picks up the the encoder or whatever and that is your home position. Then the control knows it can go *x* amount in each direction and it will go that distance or until it hits a limit switch, which ever comes first. Its old tech on an old CNC Knee mill. I don't think these were really intended for continuous production work, but rather for one off prototype stuff. But it has a CAT40 spindle and 5-7hp so it can get it done when it needs to.

[sortafast]

OK, So I have tried it many different ways, and can't seem to get it to work. I lack the mean to test the encoder in house. It's an HEDS-9140 #A00. I have tested the tach and it outputs some voltage when I spin it by hand, usually less than 1 volt max, but I can't exactly rev it up manually. There is no electrical info anywhere on the servo or on the tach to indicate what the output should be. I reversed the polarity on the tach and switched it back. Still running away. I didn't disconnect any of the encoder wiring when I originally took it apart. I checked and there is continuity from the encoder and attached pcb to the plug for the whole servo. Nothing was off there. On the PCB there is one of these http://www.mouser.com/ds/2/405/ds26ls31c-442850.pdf . Not sure what that is for, a bit above my pay grade. I can get on a scope by Wed next week, but I want to try to have this patched up before then if possible. I am thinking that the encoder isn't getting signal for some reason. It's either the vertical alignment of the disc or the encoder itself is shot. I am going to try to align the disc, again, but since I dont have the tool and this thing mounts kind of odd compared to the drawings, it might be a bit tricky.

[Al_The_Man]

Unfortunately those are rather crude differential quadrature encoders cost about $25.00.
There is info out there on the web as to their design, the IC you link to just converts the encoder signal to differential type output.
Is it possible to do the bench test for the tach I mentioned?
Al..

[sortafast]

I got the thing spinning at around 300-ish rpm or so, and it was pulling 3-5V off of the tach. So at least that seems to be working to some degree. I tried swapping the polarity, no dice. Just revs up and faults. I think the problem is somewhere in the encoder. It's either the vertical alignment of disc, I managed to get some contamination into the encoder head or there is some sort of electrical issue or fault in the encoder it's self. I kind of wonder if when I fire it up and clear the e-stop faults, that it sends a signal to the servo to spin a tiny bit to make sure it picks up the encoder, and when it is not present it just spins up then faults out once it hits a certain speed.

[sortafast]

.