Build Thread New (to me) WM30 mill. - Page 5


Page 5 of 6 FirstFirst ... 23456 LastLast
Results 49 to 60 of 65

Thread: New (to me) WM30 mill.

  1. #49
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    As it always happens, as soon as I type out that I can't find the information, I think of one more word to look for and I find the answer.

    Searching for "contactor terminal labels" got me sorted out. The A1 and A2 connections are the coil. It's a 120V coil so easy to deal with using only mains. The L1-3/T1-3 connections are meant for the three phases of an AC motor, but they are just three normally open, high current contacts the coil controls. The other NC/NO terminals are extra relay contacts controlled by the coil as well. Perfect for what I want. I'm glad Fastest1 prodded me to take a second look as that's $100 I don't need to spend.

    The wiring will look like this:

    Mains come in to DIN terminals => Breaker
    Breaker comes back to DIN and splits to => Contactor L1 *AND* ESTOP / Power Rocker Switch in series.
    ESTOP and Main Switch => A1 (coil)
    A2 (coil) goes to => Neutral

    T1 goes to the "Machine Enable" SSR and from there to the KBMM L1 (hot).

    Same arrangement for the seperate inbound mains to the stepper PS minus the split to the coil.


    The auxiliary NO relay contact is then used to signal ESTOP back to the PC. With the coil energized the ESTOP input is grounded and logic low. If the coil is not energized the input line will float high and that signals an ESTOP condition to the PC. This arrangement means pressing ESTOP shuts down all mains power past the breaker and signals ESTOP to LunuxCNC (or power loss to the machine does the same). Machine enable allows the PC to also control power via the SSR's but without generating the same condition as ESTOP would (the coil is still energized).


    Atomarc: Yes there is an ESTOP input on the MX4660. However, because it relies on the internal transistor switch in the MX4660, the drive has to remain powered for LinuxCNC to know anything is wrong via the "fault" signal. One of my design goals in this build was to make it possible to have ESTOP kill all power to any high power devices as there could be a condition unrelated to the machine operation that needs the power shut off RFN. The power supply is smoking, the motor caught fire, a pipe burst ect. My previous system did largely what the MX4660 does, ESTOP cut the enable signal to the BoB. That's not really the same thing. It's more of an abort. The interesting thing was that in so doing the steppers remained powered and locked. Here again, what if the problem is in the drive or the stepper motors themselves. Doing ESTOP like that does not really do the trick when it's just shutting down the BoB.

    I wanted two things:

    1.) ESTOP shuts down all the high power equipment (Spindle, Steppers, Coolant, Ect.) and LinuxCNC is aware of ESTOP condition still.
    2.) LinuxCNC turns the equipment on and off from Virtual Control Panel, Physical Panel, or Pendant and powering down is not the same as ESTOP.

    The contactor simplifies (1) and the SSRs handle (2). The side bonus is that if we lose power in the shop the computer is on UPS but the machine is not. When it loses power the PC will see that as an ESTOP thanks to the auxiliary relay and LinuxCNC quits trying to move. That's a frustration in my current machine because in that event if it's a quick on-off loss it can wreck the part or break an endmill when the stepper motors come back on, the spindle doesn't, and the PC never knew there was an issue.

    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  2. #50
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Well it took some doing but I was able to work the contactor into the space and had just enough terminal blocks to get the breaker wired in for each circuit. Pretty slick now. Pressing ESTOP drops the contacts to all the line powered devices on the board. LinuxCNC looses the logic zero on the ESTOP input and set's it's internal state into estop as well. That clears the machine enable signal forcing you to power on the machine again from the control panel.

    I need to make an aluminum mount for the SSRs to keep them cool and I need one more SainSmart relay module because they are made active low and that matches perfectly to the 7i92's power on state that defaults to all inputs and all signals pulled high. The default state must be inaction.

    Thanks again to Fastest1 for bringing up the contactor. That worked perfectly!

    New (to me) WM30 mill.-img_1457-jpg

    Attached Thumbnails Attached Thumbnails New (to me) WM30 mill.-img_1457-jpg  
    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  3. #51
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    A cautionary tale:

    I had been pretty happy with myself for making no wiring errors up to now. My A/C wiring was all good and pretty clean. I had the system controlling spindle on/off as well as forward and reverse correctly. Life was good. Now I needed to do what i had been dreading.... the spindle speed control. Knowing the KBIC's have that at line potential I was really uncomfortable with it but had read enough to learn that the MX4660 was supposed to have it's PWM section isolated from the rest of the electronics. With a bit of nervousness I connected that up:

    MX4660........KBMM
    +10V In.........P3
    PWM Out......P2
    EGND...........P1

    The result was no sparks or smoke and volt meter said that +10V to EGND was showing right at 10V. So far, so good. When commanding spindle speed though I was getting strange readings. The voltage seemed close to pegged out no matter what speed above a few hundred RPM I selected. I tried 1000 to 5000 with only a couple tenths of a volt different. I was not sure about what I was reading and I wanted to be certain that the pin was really just producing a clean voltage. This is the origins of my doom.

    I saw the oscilloscope sitting on the bench and thought that I would hook that up and run a program that ran through spindle speeds in a linear climb and see what the output on the scope was (probably a dead end anyway with how the S command operates). So I power down the machine and connect the scope to P2 and ground the probe to P1 to watch the output. I power back on and start the spindle then notice: "Hey, the scope is not plugged in". Some of you KBIC vets can probably guess what's about to occur. Somewhere, deep in my lizard electronics brain a dim red light comes on. I recall something about wall powered scopes and the probe ground. Can't recall what exactly though. Thought about it more but my inner Einstein overrules the lizard. "Right, I think it had to do with different ground potentials on the two scope probes or something, anyway, I can't remember but it seemed like a rare thing to watch out for and I'm only using one probe". "Onward ser Idiot, plug that sucker in and let's go".

    The pop was exceptionally loud the moment I start to slide the prongs into the socket and I found myself in darkness in the basement. Panic hits as I toss the scope cord away and look up at the controls. No clear smoke but a faint wiff of that unmistakeable odor of dispair. I pull the plugs on everything and head over to the breaker panel. The 15A breaker is tripped. "Crap.... CRAP..... CRRRRAAAAPPPPPPP", only not the real word i'm using. Then I begin to power-up each section and verify whats good. The 7i92 is still working and undamaged. I prepared myself for real heartache on the MX4660, but it also powered up and was working. The KBMM powered up but the spindle brake light was not on as it usually was. A check with the meter shows it's fuse is also blown. I replace the fuse and it powers up and the spindle brake light is lit. I'm hopeful but when I release the brake signal by turning on the spindle in Linux the light remains on. Further there is no voltage present at the P1 and P3 terminals. So I have fried something on the KBMM and it's not clear just what yet. I stood slack-jawed for a while trying to understand WTH just happened.

    Finally it dawned on me. My lizard brain was right. The d$%@n scope probe ground is connected to earth ground on the A/C plug. The moment the ground prong touched the socket the ~100V potential of the P1 pin was grounded through the scope probe. That is what lizard-brain was trying to warn me about. A line powered OScope can't be referenced to ground at potential above earth ground. I suspect at least the zener diode involved blew out and possibly more. Unfortunately for me this 125R was an EBay score that was a great deal for a working unit at just a hair over $100.00. No other such deals are out there to be found right now. Most are close to the price of a new unit direct from KB, or take weeks to deliver. So I kicked my own butt verbally for 20 minutes and ordered a new one from KB. I'll attempt to repair this unit if I can figure out what's blown up on it.

    Good news was the very expensive digital OScope is no worse for wear, as that could have been a real punch to the gut if I had killed it too.

    The really upsetting thing is that I figured out the issue with the voltage output about an hour later. The output is just highly non-linear. After mapping out the command RPM, vs the percent of max voltage it was clear the PWM output maxed out at 1000RPM despite the scale variable and was still highly lopsided with the first 30% of adjustment occurring in less than 100RPM. By mapping the real-vs-commanded percentages out I was able to create a "lincurve" device in LinuxCNC that mapped the system's requested RPM to the PWM RPM value that would actually produce that percentage of voltage. The result was a much better linear output as RPM commanded is increased. All I needed was the d$&^n voltmeter, some paper, and some time.

    Listen to the lizard-brain!!!!

    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  4. #52
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Always a neat day when you hook up stepper motors and hear them start singing to a program file! Looks like the configuration is solid enough that once I get the machine mounted on the stand I will have X/Y/Z motion ready to go. My replacement KBCC-125R arrives today but I'm out all weekend so probably next week before I hook up the spindle for real and test RPM response.



    I thought I would post up the relevant configuration files from LinuxCNC for what I have working so far. These are not likely to be that helpful for understanding how to get a very basic LinuxCNC config going as there is now a lot going on. However they do illustrate how to accomplish a few more advanced items. First a quick rundown on the files.

    WMD30-7i92.ini - This file contains most of the fill-in-the-blanks variables that LinuxCNC needs. There are sections for system defined variables as well as sections that allow user-defined variables.

    WMD30-7i92.hal - This is the HAL (Hardware Abstraction Layer) configuration file that builds connections between HAL pins and components. Most everything in LinuxCNC is a component and has certain input/output pins, and parameter values. This file determines how those pins are connected together.

    WMD30-postgui.hal - This is another HAL file with the same format as the above. It is distinct however in that it runs only AFTER the GUI has been constructed. This is key for allowing the use of virtual control panels because otherwise the pins for the widgets on the panel would not be available at the time that the normal HAL file loads.

    WMD30-virtual-panel.xml - This file describes the virtual control panel in PyVCP that loads into the Axis GUI when starting up. It's referred to in the "WMD30-7i92.ini " file. It's a tedious job to create it as nothing is visual, you have to type out the XML and then test and check the look. There is a newer method of building virtual panels called GladeVCP but I stuck with what I knew from before.


    Some things to note that burned me in my first runs with LinucCNC:

    While connecting one device to one pin is fairly easy, there is no notion in HAL of "wired-or" or light-switch-or. You can not connect two OUTPUTS to one INPUT and expect the INPUT to be true if either OUTPUT is. HAL enforces the notion that only one output can be present on any signal network. This means that if you have multiple devices that must output to a single input then you have to use a multi-input logic component to handle this. However any number of inputs can be added to a signal network.

    A pin can not be in two signal networks. This makes compartmentalizing the configuration a bit tricky. So say you have a network called system-spindle-on and that signal connects a push button input to the output pin that wires to the spindle-on relay on your BoB. You can't have another net named system-spindle-on-light that connects that push button to an LED input on the virtual panel. The LED must connect to the existing system-spindle-on network. Give some thought to how you will name signal networks because it's going to get complicated in a hurry.

    There are few wizzards for anything. You will have to do a lot of reading to understand how to 'wire' things up in HAL. For instance there are no GUI interfaces to linearizing your spindle PWM signal. You will have to do the math and concoct the proper statement to load and configure a component to get in between the raw LinuxCNC RPM setting and the actual PWM RPM setting to re-map the output. Another good example was the need to create a lookup table that took spindle-cw and spindle-ccw signals and create a single spindle-fwd-rev signal for a single relay. This look up table was designed to always set the spindle forward unless spindle-cw is 0 and spindle-ccw is 1. All other possible states result in forward as the output. Otherwise you would be stuck with two separate output signals for direction. E-Stop always seems to get more complicated than I'd like as well.

    You can only load a type of component once. If you load the two input OR gate component (or2) then you have to declare every needed instance in that load statement. You can't later load another 2 or 3 of them in your postgui.hal file.

    Error tolerance is terrible in LinuxCNC. Any errors in HAL or the XML will result in a program crash with a very dense and scary looking wall of text. However if you scroll past the first block of crash-babble text, you will see a text description of why it crashed that usually points to the file and line number. XML editing is a special case of suck, where it often only states the file failed to load without any indication of where it went tango-uniform. As a result I recommend making changes incrementally and testing or you will be tearing your hair out trying to pin down which of the 30 edits you made is really causing the problem.

    I suspect there will be further tweaks to these files after I get the machine up on the stand. None of the physical control panel items are really built and I'm debating just how I'm going to do that anyway. Also not certain where backlash and screw compensation will be until things are actually running but the meat of the configuration is there and it's a pretty solid view of how an MX4660 and 7i92 Ethernet Anything I/O can be setup together.

    Attached Files Attached Files
    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  5. #53
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Whew.

    Well it's been a long road to get here but I think the electronics are all nailed down. When compared to my current machine it's night and day different. There was no spindle control, no REAL estop, no control available for any other external devices. Just step/dir and a power switch. This is a much more robust control and I feel pretty proud of it overall. So now we are really at a point where the majority of the work ahead is mechanical.

    - I need to get the machine castors mounted to the 4" steel box tube and mount that to the machine stand. Together these will raise the cabinet about 7" to 8" off the ground.

    - Mount the Machine Cabinet electronics inside the Grizzly machine stand.

    - I need to find someone to create a metal pan for the machine that doesn't cost as much as the machine and run some plumbing for the drains.

    - Build some walls from 80/20 extrusions to make a chip/splash guard wall around the pan.

    - I need to create some interface panels for the connectors.

    The High Power Machine Cabinet Gets:
    - Steppers will use 4 pin, CB/Mic XLR style connectors to carry the A/B motor coil outputs from the MX4660.
    - The spindle will use the same, I'll run 2 pins for (+) and 2 pins for (-).
    - Molex connectors for Limits, Spindle Encoder. (I don't love these but I'm invested now. Had to use epoxy to make them hold steady in a panel mount configuration.)
    - A second panel is possible for coolant and aux.
    - Third panel for interface to machine cabinet with DB25, 2 RJ45, and Molex for the ESTOP switch.

    The Control Cabinet gets:
    - A panel for the DB-25 and 2 RJ45 passthroughs.
    - A panel to make the exhaust grill for the fan.


    Once this is done, it's largely a matter of putting the machine on the base and connecting it up. When you stop and consider the work a Tormach or Novakon seem's like a bargain. Even with these done I'll still need to design, obtain, and install a flood coolant system.

    Sigh.


    I'll try to get a video up later of the controls all working.

    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  6. #54
    Registered ninefinger's Avatar
    Join Date
    Sep 2006
    Location
    Canada
    Posts
    424
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Quote Originally Posted by photomankc View Post
    ... When you stop and consider the work a Tormach or Novakon seem's like a bargain.
    Now that I have built a machine and its still not "done" to the level of a Novakon or Tormach I whole heartedly agree. Never would have thought it when I started out - figured I could do better or at least cheaper than them by building myself. "Mass" production has its benefits in this case.
    Your build is looking good - I like the MESA products and their support for LinuxCNC - should have used them on my mill (may still convert it), I did end up getting a 7i72e for my little lathe and its very capable piece of hardware, love the ethernet connectivity (i.e. I can swap out PC's almost as easily as a stepper drive).

    Mike



  7. #55
    Registered
    Join Date
    Jan 2008
    Location
    Australia
    Posts
    906
    Downloads
    1
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Ninefinger: I'm not waste of a 7i72. Typo? 7i76e? 7i92m?

    7xCNC.com - CNC info for the minilathe (7x10, 7x12, 7x14, 7x16)


  8. #56
    Registered ninefinger's Avatar
    Join Date
    Sep 2006
    Location
    Canada
    Posts
    424
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Uhg! yes should have been 7i76e - apparently I had too many drinks before posting that... the ethernet connected card for up to 5 axis with steppers.
    Mike



  9. #57
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    Man, that's a nice bit of kit. I probably should have really gone with that for the more robust I/O. That's my biggest gripe with MESA. They have a lot of really good gear but it's hard to sort through the options and cards. The MX4660 does a lot of the job of isolating I/O for me. However, machine inputs are the hardest to come by in this setup. The fact that the 7i92 driver drops the spindle encoder inputs on the main DB25 I/O pins first robs me of 3 isolated inputs right off the bat. If I ever want a tool changer I may be missing those inputs.


    I will definitely come in less expensive than one of the big two, but it will never have the polish of those machines either. There are some things that I am improving greatly from my last design but they are still not on par with the manufactured machines. Atomarc is resetting the bar for build quality here but were I to go after that level of build I'd not be done for another 4 years given my budget, tools, and skills. Yeah you save several thousand dollars building on your own but the investment in time is quite significant.

    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  10. #58
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    I knocked out a few of the cutout cover panels I'll be using the last couple of days. Since I'll be making several, I knocked up a simple fixture from 1/4" plate so each one isn't a custom job. All the engraving and critical features are made on the top side along with profiling the outer edge after the screws are secured to the fixture. Makes producing them a little easier. All work co-ordinates are based on the fixture not the stock. The backs are relieved in depth to between 0.035" to 0.075" depending on what the items need. That I just line up by eye as nothing is really critical dimension or clearance-wise back there. The holes for the RJ-45 keystone passthroughs were the tightest margin, but they worked out nicely with a tight fit the first time, just a bit of filing in the corners needed. That saved around $13.00 since the round-hole passthroughs on Amazon were twice as expensive as those square ones I found at Lowes.

    "Master Enable" is just a switch in the E-Stop circuit line. If it's switched off then the machine is in the same state as E-Stop and can't power up. Originally I was going to put that inline with the enable signal from the controls but I like this better. The other way LinuxCNC would not be aware of the condition. This way it will keep the fault light lit, and keep all the gear, even the contactor, powered off.

    New (to me) WM30 mill.-img_1490-jpg

    New (to me) WM30 mill.-img_1492-jpg



    The machine cabinet board wiring is done now. What remains is connections to the panel connectors. The last bit of HAL file work is setting up for switch speed control between high and low gear. A MUX component will be used to route the right limit and curve correction output to the PWM input pin based on the High / Low selection (1 / 0). This could easily be made a physical switch on the mill, or even an automatic selection based on a photo sensor if I get really ambitious down the road. This way allowed minimum and maximum RPM will be tied to the gear selected and be properly represented on-screen and to the speed control PWM.

    New (to me) WM30 mill.-img_1493-jpg


    I noticed I had not posted a photo of the control cabinet with the metal bezels installed. So here we go. First picture is what it looks like with an MX4660 fault, or EStop, or power loss. The second is with all systems green and machine enabled.

    New (to me) WM30 mill.-img_1496-jpg

    New (to me) WM30 mill.-img_1495-jpg

    Attached Thumbnails Attached Thumbnails New (to me) WM30 mill.-img_1490-jpg   New (to me) WM30 mill.-img_1492-jpg   New (to me) WM30 mill.-img_1493-jpg   New (to me) WM30 mill.-img_1495-jpg  

    New (to me) WM30 mill.-img_1496-jpg  
    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  11. #59
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    I've been working on running in the spindle bearings. Started out overheating at anything above 3700 RPM. With some time spent running in from 3500RPM I've been able to work up to 4300RPM while holding below 150*F. At each speed I've seen the temp climb to around 150 and then start to drop back down slowly. I think I will run at 4500 max for a few months and then begin working on getting up to 5000. Should be a little more worn in by that point. I think 5500 is going to be a stretch,

    In the process of doing this I noticed the SSR for the spindle was getting pretty hot. The metal on the bottom was too hot to touch. I had always assumed that I would mount them to aluminum plate eventually but then I decided to make something with a bit more flair. So I made a custom heat sink with a few fins to mount the SSRs onto. I do love Fusion 360's adaptive clearing toolpath. The center slot was cut at 0.375 depth with 0.08 WOC and 50IPM. Without hard corners, there was just a steady hum and a rooster-tail of fat, hot, chips flying out of the cut. Makes me smile.

    New (to me) WM30 mill.-img_1502-jpg

    New (to me) WM30 mill.-image1-jpg

    Going to take a few days off to handle some other items then start on mounting up the interface panels on the machine cabinet.

    Attached Thumbnails Attached Thumbnails New (to me) WM30 mill.-img_1502-jpg   New (to me) WM30 mill.-image1-jpg  
    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


  12. #60
    Registered
    Join Date
    Dec 2009
    Location
    USA
    Posts
    1401
    Downloads
    0
    Uploads
    0

    Default Re: New (to me) WM30 mill.

    2 of the 4 interface panels installed in the machine cabinet - 2 more to go (AC outlets for coolant and auxillary, and input panel for Small Control Panel).

    New (to me) WM30 mill.-image1-jpg

    New (to me) WM30 mill.-img_1508-jpg



    More work on the redesign of the pan. Going to try and have a metal pan made. About 68" long, 30" wide, and 9" deep. Center of the pan will have 4" by 12" slot for coolant/chips. Plan for a drawer under the pan that will have a screen to collect chips and hose at the back to return coolant to a tank. I'm still looking at how I want to design the feet the mill will sit on to lift it over the pan and let chips/coolant flow under. I expect this will be a bit pricey, but I want metal and not fiberglass coated plywood.

    New (to me) WM30 mill.-screen-shot-2017-09-20-1-12-a



    Finally, the design of the Small Control Panel is taking shape. This is just some of the common buttons I would use for convenience rather than having to hold the pendant all the time.

    New (to me) WM30 mill.-screen-shot-2017-09-19-2-30-a


    Top -> Botton and Left -> Right:

    1.) ESTOP - Separate 120VAC connection.

    2.) Program Run

    3.) Program Pause

    4.) Program Stop

    5.) Aux On/Off

    6.) Coolant On/Off

    7.) Tool Change Complete.

    8.) ABORT!


    More.... more work to do, always more.

    Attached Thumbnails Attached Thumbnails New (to me) WM30 mill.-screen-shot-2017-09-20-1-12-a   New (to me) WM30 mill.-screen-shot-2017-09-19-2-30-a   New (to me) WM30 mill.-image1-jpg   New (to me) WM30 mill.-img_1508-jpg  

    Last edited by photomankc; Today at 08:47 AM.
    CNC: Making incorrect parts and breaking stuff, faster and with greater precision.


Page 5 of 6 FirstFirst ... 23456 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

New (to me) WM30 mill.
New (to me) WM30 mill.