Shop OBEX P1 Docs P2 Docs Learn Events
All-in-one CNC machine controller (stepper motors) - Page 2 — Parallax Forums

All-in-one CNC machine controller (stepper motors)

24

Comments

  • Oh, I didn't mean to infer that you were out of touch :D

    No, I mean the likes of Pilz, SICK, Telemec, etc. They have the industry buying-in to this BS that they paint yellow or bright red, stick a "safety" label on it and put the pricing through the roof.
    I come across machines, all the time, where the operator is standing on a "safety mat" (4-wire, fail-safe) and he is totally convinced that he couldn't be safer. In actual fact, the mat controller feeds a "Hold" input that merely sets the servos to zero velocity....but they are still enabled and holding position :o
    Meanwhile these guys are all over the machine with their hands inside hydraulic clamping dies, etc.
    I keep pushing my ZEH (zero-energy-hold) retrofit package to the management but they don't want to hear about it. ZEH blocks all forms of energy (electrical, pneumatic, hydraulic) during a hold condition.

    I have yet to meet an in-house safety officer that even understands how anything works. It justs needs to be a bright colour and everything is fine. :|

    Craig

  • It's probably not very useful (or even meaningful) to anyone but me but sometimes ASCII-art is beautiful :D

    stateChart
    '      out    duty SS1  duty ES1  duty ES2  duty SS1  duty ES1  duty ES2  duty SI1  duty SI2  timeout
    ' state #0  post
      byte %000,  h1,err1,  h1,err2,  h1,err3,  h1,err1,  h5,err2,  h5,err3,  h1,err4,  h1,err5, t05,idle
    ' state #1  idle
      byte %000,  h1,sta1,   0,nac,   h1,err6,  h1,errA,  h1,err6,  h1,wai1,  h1,err4,  h1,err5,   0,nac
    ' state #2  sta1
      byte %100,  b1,idle, h90,sta2,  h1,err6,   0,nac,    0,nac,   h5,err6,  h4,dor1,  h1,err7,   0,nac
    ' state #3  sta2
      byte %110,   0,nac,  h97,err6, h90,open,   0,nac,  b90,idle, h97,err6,   0,nac,   h4,clos, t05,err9
    ' state #4  open
      byte %111,   0,nac,  h97,err6, b90,stp2,   0,nac,  b90,stp1, h97,err6,  h4,dor1,  h4,dor2,   0,nac
    ' state #5  dor1
      byte %111,   0,nac,  h97,err6, b90,stp2,   0,nac,  b90,stp1, h97,err6,  b4,open,  h4,clos, t05,err8
    ' state 11  dor2
      byte %111,   0,nac,  h97,err6, b90,stp2,   0,nac,  b90,stp1, h97,err6,  h4,clos,  b4,open, t05,err8
    ' state #7  clos
      byte %110,   0,nac,  h97,err6, b90,stp2,   0,nac,  b90,stp1, h97,err6,  b4,dor2,  b4,dor1,   0,nac
    ' state #8  stp1
      byte %000,   0,nac,   h1,open,   0,nac,    0,nac,   h1,err6,  b1,idle,  h1,err4,   0,nac,  t05,wai1
    ' state #9  stp2
      byte %000,   0,nac,   b1,idle,  h1,err6,   0,nac,    0,nac,   h1,open,   0,nac,   h1,err5, t05,err9
    ' state #10 wai1
      byte %000,  h1,wai2,  h1,sta1,  h1,err6,  h1,errA,  h1,err6,  b1,idle,  h1,err4,  h1,err5,   0,nac
    ' state #11 wai2
      byte %000,  b1,idle,  h1,sta1,  h1,err6,   0,nac,   h1,err6,  b1,idle,  h1,err4,  h1,err5, t05,err9
    

    I wrote an interpreter that iterates through the table of the state machine. It compares the duty cycle of inputs to a threshold value and changes the state if the threshold is reached. "h97" means "higher than 97%" or "b90" means "below 90%". "nac" = no action

  • Next milestones:

    • safety state machine verified - check
    • chip select decoder and SPI driver for Trinamic ICs works - check
    • bus termination RCs optimized - check
  • Can't wait to see the end result :+1:B)

    Craig

  • evanhevanh Posts: 16,023
    edited 2023-08-03 23:26

    @Mickster said:
    Production rates weren't acceptable so the always-overruling production management had the maintenance department bypass the safety gates.
    Operator gets hurt, runs to an attorney and I end-up paying him six-figures :s (typical in USA)

    Those who come-up with "dual-redundant, forced contact relays" are out-of-touch with the real world.

    Personally, I would prefer intelligent monitoring devices that communicate directly with the control CPU using a protocol that can't be bypassed.

    What you're describing is security, not safety. Two quite different problems and two quite different legal statuses too. If someone chooses to bypass a safety then that's on them.

  • @evanh said:
    What you're describing is security, not safety. Two quite different problems and two quite different legal statuses too. If someone chooses to bypass a safety then that's on them.

    In the USA, what you wrote above is incorrect.

  • evanhevanh Posts: 16,023

    People often get the two confused. Lawers are no different.

  • @evanh said:

    ng a protocol that can't be bypassed.

    What you're describing is security, not safety. Two quite different problems and two quite different legal statuses too. If someone chooses to bypass a safety then that's on them.

    Sure, but here's the dilemma:
    The customer never entered into the equation because the injured operator wanted to remain employed. So, it's on me to point out the fact that the customer was at fault and drag him into the legal process.
    The problem is that the customer is a regular seven-figure/year account. I drag them in and the account is history. His attorney was aware of this and they call this "the good business decision". Get screwed-over for six figures or pi$$ off the seven figure account.

    @JRoark
    Sorry but this is totally correct. I had decades of this horse doo-doo. This is why you have bill-boards all over the place for scumbag lawyers who realise that all they need to do is become a PITA in some frivolous BS, just to get an out-of-court settlement....The system allows this garbage. They made such a big deal about the Mcdonalds "hot coffee" incident like it was something extraordinary....no it ain't.

    Craig

  • @Mickster
    Pssst! Dude: I’m agreeing with ya, hermano. Read it again. 👍 (and if it makes ya feel better, the “hot coffee lady” got mostly zilched-out at the court of appeals. The COA was not amused, and they came down for McDonalds pretty hard. Hehehe)

  • In E-stop mode you want to disable the servo drive power. If hydraulics counter balances a heavy Z axis load, there should be solenoid valves to lock the cylinder position when they turn off, then hydraulic pressure can be dumped to tank.

  • evanhevanh Posts: 16,023

    @Mickster said:
    Sure, but here's the dilemma:

    Yeah, probably happens way more often than most would believe.

  • MicksterMickster Posts: 2,698
    edited 2023-08-04 15:04

    @JRoark said:
    @Mickster
    Pssst! Dude: I’m agreeing with ya, hermano. Read it again. 👍 (and if it makes ya feel better, the “hot coffee lady” got mostly zilched-out at the court of appeals. The COA was not amused, and they came down for McDonalds pretty hard. Hehehe)

    My apologies....read it too quickly :)

    Craig

  • "zilched-out", "hermano", "COA", "seven figure"??? :/ I don't understand anything.

  • ErNaErNa Posts: 1,752

    It's wording needed to create 20% of GDP by legal disputes ;-)

  • MicksterMickster Posts: 2,698
    edited 2023-08-05 10:27

    :D

    Hermano = bro (brother)
    Zilch = zero, nothing (she got nothing, she was zeroed-out)
    COA = Court Of Appeals

    Craig

    Edit: six figure = hundreds of thousands
    Seven figure = millions

  • I only knew a "figure of four" :D

  • @ManAtWork said:
    I only knew a "figure of four" :D

    We had a dev who was into this same crazy stuff and it was a constant worry that one day he might not show up :#

    Craig

  • @ManAtWork said:
    I only knew a "figure of four" :D

    Appologies for the offtrack...
    AKA "unscheduled meeting with a beer truck" referring to the one dev or otherwise trained key person not available permanently.

  • @Mickster said:
    We had a dev who was into this same crazy stuff and it was a constant worry that one day he might not show up :#

    The funny thing is that there is a high chance that doing this "crazy stuff" actually increased the probability of him showing up at work. Among the top ten reasons why workers don't show up is back pain. Climbing training is one of the best preventions to avoid back pain.

    This is getting more and more off topic but a) I like talking about crazy stuff and b) it's also about risk management and can be used to make working with machines safer...

    IMHO, it's not so much what you do but how you do it that makes things safe or dangerous. If you something that you know is dangerous you are highly focused, double check your gear and do anything that makes it as safe as possible. Most accidents happen when you are in a hurry or bored and distracted or both. A women out of my home village died in a car accident because she was using her phone while driving. A friend of mine crashed his car because he fell aspleep. It was late at night, he was tired but thought he had to drive to his girlfriend "at all cost". Luckily he survived without serious injuries. That's called "get-there-itis" and can be very dangerous.

    Another case of "get-there-itis" was the only case of a climber being injured who I personally know. He thought he had to complete a hard boulder "at all cost". So he didn't pay attention to the signs of his body, ignored the pain and snapped one of his biceps tendons. But this has nothing to do with the danger of rock climbing. It can happen anywhere when you exceed the "absolute maximum ratings".

    Back to the safety of machines... I think a much better way to avoid accidents was to avoid time pressure and boring or frustrating work. But it's much easier to enforce technical safety features by law.

    A manual lathe doesn't have a saftey door. The operator stand directly in front of the rotating workpiece. But he needs to pay attention all the time while having his hands on the dials. So the chances that you get your hands into the rotating parts are relatively low although the machine seems to be quite dangerous. At a CNC machine you only change the workpiece, press a button and wait until the machine completes the cycle. You make the same movements the whole day long and don't really pay attention. This is why CNC machines need to be "fool proof".

    The only time it really needs attention is when there is a problem. Then the stress level is high because the boss is not amused when the machine doesn't produce output, say: money. The operator often needs to fix it "at all cost". Safety circuits are "battle shorted"... again: "get-there-itis".

  • @ManAtWork said:

    This is why CNC machines need to be "fool proof".

    The only time it really needs attention is when there is a problem. Then the stress level is high because the boss is not amused when the machine doesn't produce output, say: money. The operator often needs to fix it "at all cost". Safety circuits are "battle shorted"... again: "get-there-itis".

    Exactly. I received a call from a very angry customer because one of his operators grabbed the "safety cord" that ran the entire length of the machine and the machine continued to run. Upon investigation, we discovered that, at some point, their own maintenance crew had worked on the machine and they kept accidentally triggering the safety cord, so they jumpered the contacts. They eventually walked away and forgot about the jumpers. Had this been a court case with a know-nothing judge/jury, I would be blamed for not making this jumpering impossible.
    I see wireless e-stops, even wearable types, starting to appear....bring it on, I say :)

    Craig

  • evanhevanh Posts: 16,023
    edited 2023-08-07 10:36

    Health and safety done proper doesn't work that way at all. You've got some dodgy customers avoiding their responsibilities and running under the radar. Craig, you've already admitted one customer bends your arm whenever they feel like it.

  • MicksterMickster Posts: 2,698
    edited 2023-08-07 13:31

    @evanh said:
    Health and safety done proper doesn't work that way at all. You've got some dodgy customers avoiding their responsibilities and running under the radar. Craig, you've already admitted one customer bends your arm whenever they feel like it.

    Customer didn't bend my arm, it was the operator who went straight to a lawyer. He didn't file a complaint with his employer and his employer didn't even notify me.

    File a lawsuit, 40 bucks. Goes nowhere and you lose 40 bucks. You're not on the hook for my defence costs.

    It's everywhere. This is Tenneco Automotive, not sure what they are today but last I checked, they were a $1.2B corp. They do all the ISO compliance stuff and they employ "safety officers" but it means nothing except for visitors having to wear hi-viz vests, ear and eye protection.
    I went to evaluate a 24vDC circuit at John Deere and they insisted on outfitting me with this deep-sea-diver type of fireproof garb, could hardly move but get to the production line and throughput has top priority.

    Craig

    Edit: I guess Tenneco were bought-out last November

    About Tenneco
    Tenneco is one of the world's leading designers, manufacturers, and marketers of automotive products for original equipment and aftermarket customers, with full year 2021 revenues of $18 billion

  • evanhevanh Posts: 16,023
    edited 2023-08-07 13:47

    I'd call "the good business decision" a pretty clear case of arm twisting. And as you said, the person filing knew all too well how it worked. Presumably schooled.

  • @evanh said:
    I'd call "the good business decision" a pretty clear case of arm twisting. And as you said, the person filing knew all too well how it worked. Presumably schooled.

    That was the contingency based "lawyer"...they get 30%. These types only ever see the inside of a court room for DUIs and domestic abuse cases. Customer was kept completely out of the picture and here's why:

  • @Mickster said:
    Exactly. I received a call from a very angry customer because one of his operators grabbed the "safety cord" that ran the entire length of the machine and the machine continued to run. Upon investigation, we discovered that, at some point, their own maintenance crew had worked on the machine and they kept accidentally triggering the safety cord, so they jumpered the contacts. They eventually walked away and forgot about the jumpers.

    The solution for this would be to program the controller in a way that it refuses to run unless once at startup or at least once per day the Estop input is toggled. As this is not possible with the jumper installed it would force them to remove it (or at least replace it with something other, for example a key switch).

    Of course, it's very hard to completely prohibit deliberate manipulations. But at least it avoids forgetting the jumpers. And they can't say "we didn't know" when you log the warning message.

    Recently, I actually implemented the "please toggle the Estop switch" request to appear at least once after installation of the software. This ensures that the hardware initially works and is not installed with a permanent jumper.

  • evanhevanh Posts: 16,023

    @Mickster said:
    Customer was kept completely out of the picture and here's why:

    I don't see a why in that PDF.

  • @evanh said:

    @Mickster said:
    Customer was kept completely out of the picture and here's why:

    I don't see a why in that PDF.

    The relationship with Tenneco? Corporate policy was to have a minimum of two "approved vendors" which they did but I got 100% of the orders anyway. Taking a hit for that account was a no-brainer. In-fact it turned out to be a temporary hit because they had me inflate my quote for a special project to compensate. Such was the relationship. :)

    Craig

  • evanhevanh Posts: 16,023

    Okay, so a quid-pro-quo arm twist then. That was good of them.

  • @ManAtWork

    Yes, when you brought-up "suspicious states", I could've picked up the fact that the SICK laser scanner had triggered but the safety cage door was still closed. Normally not possible. :|

    Craig

  • @evanh said:
    Okay, so a quid-pro-quo arm twist then. That was good of them.

    Yup, those responsible for appropriating funding are friendly and cooperative, however, those who get bent out of shape about being dragged into a lawsuit....not so much. :)

    Craig

Sign In or Register to comment.