416 SERVO ALARM : AXIS DISCONNECT alarm on axis movement

Hi everyone, we are having issues with commissioning a vertical lathe machine with double column and would like to probe some of your experience and knowledge.

Axis are as follows: one column being X,Z (HEAD1) and the other U,W (HEAD2). There is a rotation axis C and a crossrail Y. PMC & NC name is Fanuc F16 RB3.

The double columns axis X, U (horizontal) and Z, W (vertical) are identical. Horizontal axis X, U are mirrored with negative (-) movement toward the center of the lathe. Position is monitored via linear incremental scale heidenhain LB 326, scanning unit AE LB-326. Drivers (a06b 6078 h106) and motors are of the same type (for X,U). Incremental scale is used with the two vertical axis Z,W but don’t know the precise model or scanning unit. Drivers and motors are of the same type for Z,W. So basically this four column axis could be interchangeable between them.

To my knowledge we have connected everything correctly and ofc rechecked multiple times. When we started with the testing of the lathe at first there were no alarms – both heads X,Z and U,W axis were working as intended. After some time the error “416 SERVO ALARM : U AXIS DISCONNECT” started to sporadically show up on U axis movement and now it has progressed so that on any movement of the U axis the same error occurs.

There is one more error on HEAD2 (U,W axis) “414 SERVO ALARM : W AXIS DETECT ERR” which still shows up sporadically with no particular order. The error sometimes shows even if whole system is just ON and resting for some time. But I would like to first debug the 416 U axis disconnect error which occurs only when there is movement in U.
I am mentioning this error on W axis because both (and only) errors are on HEAD2 (U&W). HEAD1 works with no problems.

So browsing the internet and forums we have tried some suggestions but with no success.

First we instructed the system to ignore the scale completely by setting 1815 register and others. With scale off the U axis worked with no errors.
So we figured that it has sth to do with scale itself or connection. We swapped scanning unit with that of X axis from HEAD1, the error stull persists on U, HEAD2.
We have replaced the cable but no change.
We have interchanged signal inputs from U and X (connectors JV1, ENC1, SCALE1) on the central unit so that U was represented (connected) as X in the system but the axis disconnect error remains (this should mean that the unit inputs are OK, no cold joints or bad boards).

Is there anything else we can try to get to the root of the problem?

On interesting thing is that if we use cable for scale from U on X the direction of X scale switches – previously positive movement is now incremented as negative on X and thus we get excess error. Is there any parameter to set the scale increment/decrement direction?
How is that done on fanuc system?

Any help appreciated!
Thank you for your time!

What U see on the displays on the servo amplifiers?
Alarm says that NC unit has no connection with amplifier, this can be caused by malfunction of amplifier, (there is low voltage line inside that supply the linears and comunication) or linear encoder (short circuit will cause overcurrent in LV supply).

Quote:

---

Originally Posted by steryd

What U see on the displays on the servo amplifiers?
Alarm says that NC unit has no connection with amplifier, this can be caused by malfunction of amplifier, (there is low voltage line inside that supply the linears and comunication) or linear encoder (short circuit will cause overcurrent in LV supply).

---

Thanks for the reply!
Amplifiers display "0" at the start but after the error U amplifier (and W, so both HEAD2) display "-" (no error number).
You think this could be because of the amplifier? But when we set the system to ignore the linear scale input the error does not occur, shouldn't that still be the case with bad amplifier?

- means "not ready" it occures eg. When U press E-stop.
It may be caused by PSU or PLC unit. PSU does not only cover power supply, but also comunication between NC, and other parts of the system, and also between measuring units and servo amplifiers. So the problem can also be what U first mentioned linear encoder.

Our boards are:
a16b 1212 0871 (PSU BI)
a16b 3200 0010 (FS16B MAIN)
a16b 2202 0400 (OPT2)
The HEAD2 (U&W) are connected to the 3. board, connections JV5, ENC5, SCALE 5 for U and 6 for W.
By PSU, do you mean board 1. or the PSU for the motor drivers (unit a06b 6087 h126)?

Is it possible to switch the direction of linear encoder increments by program parameters?
If we use linear scale cable from U on X, the direction of movement and scale "increments" are opposite (and we get excess error). So we cant rule out the cable completely because cant proper test it on X.

So today we tested some more combinations:
1. We have connected X and U scales with new cable and the “416 SERVO ALARM : U AXIS DISCONNECT” still occurs. The X is OK with this new cable.

2. Diagnostic registers during the 416 error show: DGN200=00000010 (FBA is active) and DGN201=0 (none active). The documentation states this is “Pulse coder disconnec-tion (software)” and “This alarm is related with full–closed system” (see image below).

3. We flipped the heidenhain scanning unit (X, U, axis), the cables from scanning unit to EXE 602 E box, cables from EXE box to the control unit and we flipped the EXE boxes (X, U axis). We did all the combinations but the error stayed on U axis. Probably this means that the connection from scanning unit to controller (SCALE 5) is OK as the X is not affected even with all the changes.

4. Then we flipped sides by connecting U side to X inputs and X side to U inputs. So, the U side uses X inputs and amplifiers - JV5, ENC5, SCALE 5 (U side cables) connected to JV1, ENC1, SCALE 1 (X input on main board). With this arrangement the 416 ERROR flips to X axis. The X cables, connected to U inputs and using U amplifiers were OK and working. We did the 3. point again with no effects.
I suppose this means that the inputs to control unit (main board + OP board) are OK because the error does not stay on U input (JV5, ENC5, SCALE 5) but rather flips to X inputs where the U side is now connected.

5. During the testing we have noticed that the same error 416 with the same DGN200 and DGN201 values (as in point 2.) does occur when the scanning units are not physically fastened to axis U (HEAD 2) and we move either the scanning unit or axis. Electrical connections are connected. The same applies to X (HEAD 1) ofc 416 error then occurs for X axis.
So, the error is also generated when there is relative movement between scanning unit and axis (head) for example when they are not physically attached.
Could this mean there is some big gap/lag when U axis moves and linear scale signal? Is there any parameter that I could set to allow for bigger errors between movements and feedback signal? Or maybe any method to test the linear scale, servo encoder etc.?


My thinking now is either is the scale itself (dirty scale tape?), stiffens of scanning unit mount (but feels no different than X axis side) or sth on control unit (bad boards) but unlikely because of tests in point 4. with flipped U & X.
There could be some program parameters that must be set but which. Any other ideas what could cause “hysteresis” described in 5. or in general?


I would really appreciate any advice as we have already exhausted many of our ideas.

Set parameter 2003 bit 1. Uses the parameter for the soft disconnection alarm detection level.

Set parameter 2064 Soft disconnection alarm level... Fanuc presets this to 4. It's not looked at until you set parameter 2003. Depending which book you look at this is a representation of how far the axis can travel before setting the alarm. One book says each number is 1/8 turn of the motor. ( Alarm issued for a 1/8 rotation of the motor Increase this value).

Thank you for the quick reply!
Tested this today and it seems this was the problem all along. The 2003#1 bit was already set to 1 and parameter 2064 was set to default 4. We had to increase it beyond 16 to get rid of the 416 servo alarm. It appears the machine got a bit loose during transport or sth. Now it works but it needs some mechanical works/inspection.

These parameters are explained in documentation B-65270EN/05 (parameter manual) and the docs defines some additional actions for "Disconnection alarms" (image). A bit confusing since the documentation for Fanuc CNC Series 16 (B–62445E/03) specifically mentions ERROR 416 but does not reference any other actions just some vague guidelines. Thanks for your solution!

What should be acceptable value of 2064?
Does the default value 4 means 1/8 of a turn? So, a value of 1 is 1/32 and value of 16 is 1/2 of a turn?

Best Regards!

As far as a reasonable value for 2064 goes, it varies from Fanuc manual to Fanuc manual. I have customer's tell me that a loose ball nut caused their issue while another customer found a loose scale reader head. Just check out everything to make sure that you don't have bigger issue that you will be masking by setting this value too high.