17.xx control bug

I have 17 series mill software, and I have this code that does something funny I haven't figured out.

G187E0.15
G0X1.6Y-1.65(Rapid here)
G2X0.875Y-2.375I-0.725J0.F85.(90 turn, control messes up on transition to next move)
G1X-0.875(**** control messes up here, there is a hearable bump and control overshoots or something)
G2X-1.6Y-1.65I0.J0.725
G1Y1.65F300.
G2X-0.875Y2.375I0.725J0.F85.
G1X0.875
G2X1.6Y1.65I0.J-0.725

If I change my code to
G187E0.15
G1X1.6Y-1.65F300.(Fast feed here)
G2X0.875Y-2.375I-0.725J0.F85.(90 turn)
G1X-0.875(No bump this way)

This seems puzzling to me, I don't understand how a prior rapid can affect the following move.

I just ran your code samples on my VF2 running 17.2041 software and there's no detectable "bump". Both snippets of code essentially run the same.

The G187 E.15 command should essentially make the machine function as in exact stop mode.

Is your machine a next generation control? The engineers are tinkering with the software every day, it might be a good idea to get an update.

Maybe since the G187 it is trying to corner round at the transition between rapid and the feed move. That bang sound like it is just changing directions instantly? Like if you are jogging and select another axis while moving, it stops one instantly and changes to other. That what it sounds like?

Have you tried without the G187? Do you still get the bang?

My machine is a maincon upgrade to a 96 vf2, so its not next generation.

G187 E.15 sets corner rounding to something other than the standard. G187 cancels.

My software says M17.04B, and mocon 17.28

I tried the snippet I posted and it ran fine, so I added more of the prior code until this reproduces the problem. It basically acts like a rapid to feed transition. I also tried cancelling G187 which didn't make any difference.

Here is the code that does it on my machine.

%
O01000 (Test control)

G90G17
G20
G53G0Z0.
G187
G54

G187E0.2
G17
G0X1.6875Y-1.65
G2X0.875Y-2.4625I-0.8125J0.F85.
G1X-0.875
G2X-1.6875Y-1.65I0.J0.8125
G0Y1.65
G2X-0.875Y2.4625I0.8125J0.F85.
G1X0.875
G2X1.6875Y1.65I0.J-0.8125

G187
G0X1.6Y-1.65
G17G2X0.875Y-2.375I-0.725J0.F85.(Bump happens at transition to next)
G1X-0.875
G2X-1.6Y-1.65I0.J0.725
G1Y1.65F300.
G2X-0.875Y2.375I0.725J0.F85.
G1X0.875
G2X1.6Y1.65I0.J-0.725

M30

%

Do you have the HSM option enabled? If not that might be it?

friesen said:

My machine is a maincon upgrade ...

Click to expand...

Can't help you otherwise, but that right there is where your problem possibly lies!
UPGRADE!


I have machines with '01 lathe, '01 mill, '03 lathe and '06 mill controls,
They are FALWLESS!!!
Not a single mishap ever, no matter what. They either don't like it from the get-go or run it exactly forever.

Then, I've got a 2016 NextGen controlled machine.
Yes, I understand that yours isn't an NG, but the so called "upgrade" might have been produced by the same fucking college ( or worse, Google ) dropouts
who write the current NfG garbage!

I used to swear that there was nothing that a Haas control can do wrong once it's inside of it, no matter what.

Nowadays however ....

( Rant On )

My fancy NextGen control has seen 5 ( yes that is Five ) upgrades ( Read: FIXES ) in 18 months, and it is still a piece of shit compared to
the 5, 10, 15, 20 year old control!

A bit over a year ago I have yelled at D.D. Machine ( PM member ) for ranting at Haas every time he got a chance to do so about the NextGen garbage.
Well, now 18 months later, I am almost at that point!

While this may or may not have any relation to the OP, but here it is ( yupp, it is a RANT!!! ) :

The current group of assholes at Haas Automation ( aka HaasCNC ) who has any direct knowledge, job description, access, responsibility or in any way
connected to the fuckheads that write code on the new ( or upgrades to the old as it may ) controls should not be fired, rather fired upon!

Really Morons!!!
In the last 18 months I have lost more than 1K in tools running programs which ran on a '06 control for years!
Approx that much lost in tools just because the JOG was sporadically fucked up ( though you did fix that about a year ago )

BUT!

You fucking shitheads KNEW THEN, and KNOW NOW that the M97 Pnnnn call for a G8X cycle is fucked up!!!
You knew that the P value in the G8X cycle is overridden when an M97 is used for hole locations!
The P value in the G8X is some(fucking)how overwritten by the M97 P call!
You dickheads KNEW that! Knew that back in September of 2016! And yet? It still ain't fixed!

Do you want me to go on a bit further ( directed at the children coders at Haas Automation ) ?
You want me to send you a picture of a part that was destroyed by a fucked up G86 cycle that was manually interrupted and chips cleared by spindle reversal?


(Rant temporarily over )

I used to hate Google programmers (because nowadays they suck ).
I now hate Google because the ones that did not make the cut (there) ended up at Haas and are actively fucking shit up and don't know any better!

HSM isn't the cureall. This should be smooth with it off. This upgrade uses existing stuff, I don't think there is special firmware or hardware. I really like it and would not go back.

As to the ngc, it is too bad the users have to deal with this. Pretty much the way this gets fixed is by the well drying up.

friesen said:

HSM isn't the cureall. This should be smooth with it off. This upgrade uses existing stuff, I don't think there is special firmware or hardware. I really like it and would not go back.

Click to expand...

Could it be the backlash compensation that you hear? I am not sure how well the original backlash compensation values carry over to the 17.xxx software.


Sent from my iPhone using Tapatalk

Oddly enough, anything G187 causes this. I suppose it has to do with resetting the pre lookup, who knows.

What exact version of software do you have after the upgrade? I can look and see what they did on mine after the maincon upgrade.

My software says M17.04B, and mocon 17.28

I think I have an issue yet though, and I'd like params from someone else to review. When he first installed it I could hear straightline feed segments transitions, so I had to fiddle with different constant params to get it better.

friesen said:

My software says M17.04B, and mocon 17.28

I think I have an issue yet though, and I'd like params from someone else to review. When he first installed it I could hear straightline feed segments transitions, so I had to fiddle with different constant params to get it better.

Click to expand...

Search for my post here on PM for the maincon upgrade. I have all of my settings listed in that thread. Could be a FIR filter or notch filter setting. Does your machine have brushless servos?

Advances in Motion Control | Customer Resource Center


Sent from my iPhone using Tapatalk

It's brushless. The upgrade can only happen with brushless I think.

I did find here that I had set exact stop mode change to 0, which for some reason causes this.

With it off, it is quicker on transitions from rapids to feeds because it doesn't have to come to a complete stop.

Just ran the code on a coldfire UMC-750 with software 18.30A and mocon 17.28. The only detectable "bump" is at the feed to rapid transition, but not at the mark designated in the code.

Did you figure out what your problem was?

It totally goes away with exact stop mode set to 1. It always pauses at transitions now, but that's the factory settings.

Here are a list of known bugs to me at least.

#1. If the door gets opened at the wrong time or the door switch is flaky during a tool change, you'll get alarm 125 shuttle fault.

#2. Related to #1, I had a situation with the shuttle in and tool clamped in position, and the recover options would not pull shuttle back, but told me to remove tool even though shuttle was in and tool clamped. To recover I had to manually move and push shuttle to the home position.

#3. If the door gets opened and shut at the wrong time during power up restart, the axis being homed will home to the wrong place.


List of known things that should be changed.

#1. Most probing items in the menu should be changeable with "ENTER" rather than forcing one to use "F1" half the time and "ENTER" the other half.

#2. Offsets should be scrollable so that I can see G154 P1-6 rather than 1-3 and the next page 4-6

#3. Usb menu is ordered somewhat by date, or some other ordering algorithm, should be selectable perhaps.

#4. Usb takes about 10 seconds to connect, should take about 1.

#5. The ability to override %25 with door open tool changes. (This is a safety issue I suppose, but I'm not sure how it helps, as the shuttle moves in too fast to get out of the way anyway)

#6. Pressing handle jog does not reset the "DIST TO GO" like the old controls did. This is handy for manual operations.