Page 1 of 1
G31 on a Fanuc
Posted: Sat 26 Aug , 2006 20:30 pm
by Triac whizz
Anybody got any clues on this?
I'm writing a probing macro but when I make it go -X and the probe contacts I get 930 system error. +X +Y -Y are not a problem.
If the probe doesn't contact while doing a G31 -X it's happy.
Otherwise the machine works fine
I wouldn't be to bothered if I hadn't just broght a probe

Posted: Sun 27 Aug , 2006 17:04 pm
by Denford Admin
Have you got the description for the 930 error ?
As usual with Fanuc - it will be a bit cryptic - but may give me a clue
Posted: Sun 27 Aug , 2006 19:21 pm
by Triac whizz
Yup, here's what it says in the book..
CPU error (abnormal interupt). Replace master PCB
I'd believe it if the error came up on every axis.
After the error on screen it dumps a load of (what I guess is) hex - 32 bit as it's in groups of 4 but that might as well be japanese!
I'll give the chips on the memory board a thump when I take that out to add the memory
Posted: Sun 27 Aug , 2006 22:46 pm
by Denford Admin
hmmm ..... sounds like the Fanuc is crashing for some reason when probing in that direction - I'm sure there isn't a problem with the hardware but with the setup.
Abnormal interrupt is telling us that the control doesn't know where to go or what to do when the probe input is triggered.
Its years since I've setup probing on an OM-C and can't remember whats involved but I'm guessing that its a configuration error either in parameters, the PLC or your macro - not much help I know
I'd scan all the manuals you have for any bits on probing and double check every possible parameter.
Posted: Mon 28 Aug , 2006 8:40 am
by Triac whizz
uh hu, hmm what I thought as well
It can't be the macro as all 4 moves are the same just registering the interupt position in a different variable, I even tried taking out the registering of the variable, but keep turning the control off to reset and datum again is a PITA.
The macro is nothing complicated, It's just to prove it works at the moment

then I was going to embroider it...
I'll re-read (again) the manuals and do some experimenting.
So you don't think it's a Denford thing thats screwing it up, I mean have you enabled probing on other Triacs?
I know you use M29 for DNC which screws up rigid tapping (but can be got around) but then ther's no encoder anyway.. but thats another story..
Posted: Mon 28 Aug , 2006 10:37 am
by Denford Admin
I'll have to have a look at the connection manual (later this week) to see if its something we may/may not have done in the PLC
I'm pretty sure that *SKP signal (I/O or a PLC flag? I can't remember ) is something to do with it, and that you have to do something with it in the PLC
PS, can you see the PLC (PMC) ladder on your display ? some OMCs would let you see a live graphic ladder
Posted: Mon 28 Aug , 2006 10:51 am
by Triac whizz
I can see the ladder on my control, not that I entirely understand it, seeing it is all I can do, the soft keys say I can search etc but it won't let me. I guess that's another parameter I haven't discovered yet
I would have thought that since the skip signal works correctly on 3 moves it wouldn't be that, unless, of course they all go to different places?
wierder and wierder...
Posted: Mon 28 Aug , 2006 23:50 pm
by Mr Magoo
The SKIP input is hard coded to a Fanuc Input Xn.n and can't be re-assigned (I can't remember which). I guess you've worked out which one Whizz becasue you've got the Y axis and X+ working with it OK (Did we used to bring the skip input signal to a terminal connection in the electrical panel?)
To search for an input/output etc in the ladder screen I think you enter the address first, then press the [SEARCH] softkey, so enter...
X1.4 [SEARCH]
will search for an occurance of input X1.4
Yo can search for the skip input and see if it is used in the PLC - I don't expect it to be used anywhere because G31 reads the signal direct and not via the PLC (reading via the PLC would be too slow).
Posted: Mon 28 Aug , 2006 23:55 pm
by Mr Magoo
I agree that if Y and X+ works OK I don't think it is a "Main Board" hardware problem.
As another thought, could it be a problem with servo lag or similar gain/offset settings? I think your machine will have digital drives so this shouldn't be an issue but worth investigating.
Take a look at the servo lag values (sometimes called servo error - again, I can't remember which these diagnostic values are but Admin may be able to help. On the 0MC controls I think each axis had it's own page showing servo diagnostic values).
With the X Axis stationary this value should be hovering around zero. When you jog the X axis positive at a fixed feedrate you should get a certain lag value, jogging -ve at the same feedrate & you should get a similar lag value.
These lag values may will jump during axis acceleration, but should settle down to somewhere near zero again when the axis are moving at a constant velocity.
Posted: Tue 29 Aug , 2006 16:45 pm
by Denford Admin
I've found a small bit of info:
the SKIP signal is dedicated to input X8.7 (conn M18 - pin 49) but I guess you know that
You can check the status of this input in diagnostics number 8 (the left bit should go on and off when you hit the probe)
It says that the signal must be at least 10mSecs width - I guess if you have a bouncy or noisy switch, it could be switching quicker than that - that may cause a problem as the interrupt code could be called again whilst it is still running the previous interrupt handler.
Like Magoo, I also wondered about axis error - you can check these in (I think! ) the last diagnostics page in the 720 region.....
(You will see the axis following error ones flickering and continually changing value)
Just for your (or anyones) info:
Fanuc axis Diagnostic no 720-723 are axis error bits - one for each axis up to 4:
bit 0 (right hand side): overflow alarm (OFAL)
1: wire disconnection alarm (FBAL)
2: regen discharge unit alarm (DCAL)
3: overvoltage alarm (HVAL)
4: an improper current alarm (HCAL)
5: excessive current alarm (OVC)
6: voltage shortage alarm (LV)
7: overload alarm (OVL)
If you get a 404, 414, 424 or 434 alarm error, then look in the above bits where 404 alarm is X, 414 is Y etc.. to see whats caused it (normally a fuse or a breaker)
Posted: Tue 29 Aug , 2006 18:30 pm
by Triac whizz
Hmm....
The saga goes on...
The servo offset seems to be pretty much the same in all 4 directions, lagging by about 140 widgets at 100mm/min constant speed. but the servo gain shows more on the troublesome -X I'm sure I saw more than 4500 at one time (hard to tell as it changes so quick) that's against the set value of 3000 in the parameters. Could that be the problem?
Yes, I've just looked at the signal getting to the control today, yup it's there, and yes thanks to you for bringing it out to terminal connection
Axis diags at 720-2 all ok - all 0's no flickering - oops didn't look while jogging..
I hadn't thought about the switch, though since it's always on the same axis direction move, I would have thought it's very unlikely.
Main suspect must be the servo but I don't know how to go about correcting it.
Things I did notice that seemed different than I expected in the parameters were the Motors, They're Fanuc 2-0S (whats the s for sport? - slow?) I guess that they're different than a 2-0 as parameters 8*20 have 37 stored in them rather than a 6 that I was expecting (as per Fanuc manual) but then my manual doesn't list 2-0S.. another is parameter 8*21 - Load inertia ratio, mine was set at 0.
Well I've hacked around with these and it don't make diddly squ*t difference.
If you could give me a dummies explanation of the servo gain - in simple words it'd be appreciated, even if it doesn't solve the problem.
Oh and thanks Mr. Magoo - I'd been pressing the search button first didn't occur to do it the other way

Posted: Tue 29 Aug , 2006 21:44 pm
by Denford Admin
I hadn't thought about the switch, though since it's always on the same axis direction move, I would have thought it's very unlikely.
- depends on how the switch works - if its like some kind of leaf contact pressing on or off a copper ring, it could be something like corrosion in that direction - a quick test to rule it out would be to spin it around - just a thought.
Idiots guide to servo gain ()
Servo Gain - 3000 is the default setting on these controls - gives quite a tight velocity (position as well

) loop - 1667 will make a much "slacker" system, 5000 will be much tighter and will probably start the motors buzzing.
Its a magic number that has an effect on how the control responds to changes in the motors speed - the higher the gain, the quicker the control responds when you accelerate and decelerate the motor. In other words, the control is more bothered about being out of position, or at the wrong speed.
If you lower the gain, then the control is less bothered and will
lag behind the desired speed or position by a greater amount
High gain is not always a good thing for machining, as it tends to oscillate the motors - too low a gain is also not good as the axes will be sloppy at positioning / contouring
Lag is whats being shown on one of those diagnostics - its how far the actual position differs from the desired - you could probably spin one of your 2.0S motors by hand and get the lag (or error) display to increase - let go and the control will bring it back to position and show a small error ( less than 20 widgets ) - Widgets are microns or half microns if I remember correctly
Sorry - I can't remember if the parameter your looking at is the position loop gain, or the velocity loop gain
PS, you should always think back to front when using Japanese software

Posted: Tue 29 Aug , 2006 21:57 pm
by Triac whizz
depends on how the switch works
ah, uh hum, well actually it's a microswitch grabbed out of a mouse
Just to test the principles you understand...
I've been told probes don't survive G0 impacts
Looks like sourcing another board to see if that really is the problem
thanks for the idiots guide!
Posted: Sat 02 Dec , 2006 19:58 pm
by Triac whizz
Update......
I put the extra memory in & forgot about probing till recently, and tried again & b*gger me it works, so one happy bunny here and super quick G54 offset inputs cor how do you lot put up with manual probing
and no I didn't change the main board
Posted: Mon 04 Dec , 2006 11:32 am
by Denford Admin
Weird
I can only guess your old memory was a bit dubious then - if thats all you've changed ?
Posted: Mon 04 Dec , 2006 11:58 am
by Triac whizz
Well I would have thought that, but I've not removed any, only added some

Posted: Mon 04 Dec , 2006 12:01 pm
by Denford Admin
well, as they say:
If it works, don't touch it
Posted: Mon 04 Dec , 2006 12:11 pm
by Triac whizz
and If it ain't broke don't mend it
and why does hindsight always come too late ?