Best practice for mill VFD control wiring

I am back from my RV trip. (which went very well)
This post is in regards to that Arboga EM825 mill that I am working
on. It is three phase and works well. This is a 1.1 HP, 3 phase mill/drill with geared drive. It does not have and has never have a reverse switch of any kind.
My plan is to convert it to single phase by means of adding a VFD, do a nice job at that, and sell it to single phase people.
I use a Toshiba VF-A3 drive, the same one as what Pete C and Ivan Vegvary have.
So far, I got to the point of having the drive mounted and I use panel controls to run the mill. So now it does run great from single phase. This is not the most convenient and not the safest option due to where the drive is mounted (low height).
I would like feedback about the following plan:
* Mount drive out of the way to the right of the base. * Use the original START/STOP buttons on the right side for START and STOP control of VFD. * Add a panel on the left side, composed of o SPDT On-On switch to switch FORWARD and REVERSE o Potentiometer to adjust mill speed from 0 to 80 Hz. (the higher limit is the VFD setting). * Program acceleration time to be 1 second and deceleration time to be 0.4 seconds.
Is that plan basically sufficient and does it constitute "best practices"?
My own reasoning is that START/STOP is basically a safety function and it would be good to retain it in its original design for legal reasons.
i
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

If you're really concerned about legal reasons don't alter the controls. Stop buttons normally hard interrupt a contactor, a VFD control is a soft control. We had one piece of equipment, a portable floor grinder, which had a VFD and a stop button, the stop button actually killed all of the power, including the VFD. The design you are describing is an improvement over using the controls on the VFD and sounds fine for your own home use.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

OK, then, I can alter my plan:
* Original START/STOP buttons would control incoming power to VFD * The operator control panel would include: o speed control pot o On-Off-On switch to control FORWARD/STOP/REVERSE
This way, the only use for the original buttons would be some sort of emergency, but the STOP button ensures safe and complete shutdown regardless of any wiring or programming mistakes.
At the same, normally, the on/off/on switch would be used to run and stop the mill.
Any thoughts? Bruce Bergman?
i
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:

Here's what we have at work. A fused disconnect feeding the VFD and an E-stop circuit killing power after the drive to the motor, auxiliary contacts on the contactor between VFD and motor go to drive enable. The normal stop/start buttons would wire to inputs to the VFD, a selector switch to select forward/reverse, you could add a potentiometer for speed control. This way, drive gets to control the stop under normal circumstances and motor is instantly disconnected when E-Stop is pressed. You could add additional contacts to kill power to the VFD if you wanted to but most of the time the VFD will stop the spindle quicker with the power on then just killing power to it. Note that if you remove power before the VFD you're not killing power to the motor until the VFD capacitors discharge.
On some machines we use a timing relay to keep power on the drive for a couple of seconds while stopping the motor, the motor stops much quicker with the drive then it does by just removing power. We have some drives that can have a separate 24V DC logic power, you can kill power to the drive and not lose the power to the drives computer. This is done on our ProfiBus drives that are on the PLC's network.
Our drives are in a control cabinet and power coming out of the cabinet is removed in E-stop condition except control power. Those that claim you have to kill everything on an E-stop don't kill the incoming power from their utility pole, so even they don't kill "everything". What I described is what works in a ~35 year old 60+ acre factory with probably over 1000 VFD's and an electric bill of over $1 million per month. If there were ever a safety incident power would be required to be removed before the drives. Some of our techs have worked where they kill power to the whole machine on E-stop and it makes people very reluctant to use the E-stop because of the uncontrolled crash and difficulty restarting.
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
    [ ... ]

    Though if the motor is running at fairly high power, that will be pretty quick anyway.
    [ ... ]

    This is what makes me reluctant to use the E-stop on my little Emco-Maier Compact-5/CNC lathe (it's only a 5" swing, so there is not too much danger when something *does* go wrong). You hit the E-stop, and not only does it dump power from the spindle motor and the stepper motors -- it also dumps it from the computer. If I've just spent some time keying in a program and not saved it since several changes, I lose that too -- and I have to redefine the zeros for the axes -- unless I am lucky enough to have it at the home position (defined as where it is when you start the program). There are no reference limit switches to allow it to restore to a known point on its own.
    Enjoy,         DoN.
--
Email: < snipped-for-privacy@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| Voice (all times): (703) 938-4564

What I've always preferred is keeping control and input power on but killing output power on E-Stop. That way you keep your controls but disconnect the potential for movement. For example with your Compact 5 I would keep power to the computer and encoders, if it uses steppers perhaps chop the current in half or so. Kill power to the spindle motors and servo drives but not their feedback devices. That way an E-stop is going to stop motion and keep position, program, etc..
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

    [ ... ]

    [ ... ]

    Unfortunately -- the power for the spindle motor, the steppers and the computer come from the same power supply -- and it is all shut off by the E-stop button. To restore power you have to first release the E-stop button (by twisting the mushroom), and then use the key to switch the AC power off and back on.
    And -- there is not really room in there to add separate circuitry with relays and such to selectively power down the steppers and the spindle motor.
    And -- since it uses stepper motors instead of servos, drop power to the steppers and spindle but keep power on the computer and you get another problem. There is no position feedback -- it trusts that the steppers have gone to where they were told to do, so you have no idea how far out of sync the actual position of the motors and the computer's idea of where they are may be.
    And there is not really room to expand the program to allow a graceful stop to the CPU. It is running on a 6502 (64 K address space), and has about half of the address space taken by the firmware, and the other half for user programs. And the system's program is in machine code (of course, given how little space there is), and totally undocumented, unless someone has a chance to get into Emco's history files. :-)
    Some of these days, I'll take a dead spare that I have and built a servo based controller around it using EMC, so I'll have a lot more control over things.
    Enjoy,         DoN.
--
Email: < snipped-for-privacy@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| Voice (all times): (703) 938-4564

Something that might be a possibility for you. If you found the signals to your stepper motor drives you might be able to use a separate PC with a parallel port breakout board. Sort of disconnect the EMCO PC and use your own PC running EMC or similar. Perhaps you already have home switches but it would help me if I had home switches so I could set a reference point and position my cutting tools from a known position. Right now I just jog and take a cut, measure, set the axis to the radius.
My lathe is an old Anilam 14 X 20 lathe that came with good hardware and electronics except the controls were bad. I put a PC running EMC2 and I like it. My first test run I cut the EMC lathe pawn program cutting a stainless steel pawn in 3 minutes and I don't have the coolant running yet. Stainless was coming off red hot so I don't think I can speed it up without burning tooling until I get the coolant pump running.
I found the original spindle encoder was bad so I ordered one from Automationdirect.com and it cuts threads very nice now. The spindle motor is 7.5HP and I have single phase here, so I put on a 10HP Hitachi VFD and am using EMC to control my spindle speeds with RPM feedback from the encoder. I still need to wire the limit and home switches, the coolant pump, and the gearbox lube circulation / cooling pump.
I have an Anilam CNC Bridgeport mill that uses DRO scales and servo motors that I hope to be my next EMC conversion. I called Anilam about an I/O option for the controls and it was going to cost more for simple spindle ON/OFF then an entire EMC conversion will cost. But with the EMC conversion I get spindle speed, spindle forward/reverse, and can add a spindle encoder to get rigid tapping. And there's a possibility to add a 4th axis...
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

    [ ... ]

    The stepper motor drives plug directly into the CPU board, and derive power from it.

    No PC there (if you mean personal computer, instead of printed circuit). It is a large printed circuit board with the keypad, a several digit LED display, other keys for jogging and other positioning, outputs to control a tool turret (which also has no absolute home position, and the commands in the programs are "advance three stations" instead of "go to station 5", so it gets lost too when a program is interrupted (or if you forget to do enough advances at the end of the program to return to station 1. :-)

    No such luck. Nor is there much room to add such.

    Note that home switches are not accurate enough to reliably set zero from them. Typically, it is necessary to have linear encoders or rotary encoders with an index position as well as the increment encoders. So -- with linear encoders, you move until you encounter the index position, and then keep going until you are at a count zero point within that zone. With rotary encoders it is similar, except that you need a mechanical switch to say "this time around is when you pay attention to the index spot". :-)

    That is big enough to have room to hang anything you need on it. And Anilam controls were nice. I worked with one on a Bridgeport clone a few years before I retired. I found some intersting things that you could do with an Anilam and a PC to dump the entire set of canned cycles to. This allowed me to write canned cycles to engrave letters using the Hershey letter sets.

    Hmm ... if you're using clamped carbide inserts, they will happily cut with red hot chips. The only thing to worry about is whether the holder is getting hot enough to give.
    But you certainly don't want to hit a carbide insert with coolant after it gets that hot -- it is likely to crack.

    Great!
    Does it work as it stands? You certainly will have a lot of working hardware, including servo motors already mounted, even if you do go to EMC2. What interface are you planning to use -- driving the existing servo amplifiers with analog signals from a card like the Servo To Go cards, or going with the drivers which make the servo motors look like steppers to the controller? My preference is the former, once I finish converting the Bridgeport BOSS-3 to use servos instead of steppers. (I really *don't* like steppers for many reason.)

    Ouch!
    Or more than a 4th axis -- say an index/dividing head on top of a rotary table to get more creative angles. The Servo-To-Go card will drive 8 axes, though I think that there was some kind of problem which limited it to six axes with the earlier EMC at least.
    Anyway -- the Bridgeport is the first target for EMC, then the *spare* Compact-5 with the dead electronics to replace with EMC and servo motors in place of steppers. This also means a small VFD for the spindle motor so I get speed control too. Right now, it is a DC spindle motor which is selected on or off, and the speed set by a front panel pot only. I had considered adding a small relay in there and a second front panel pot, so I could switch between two speeds in the middle of a program. The primary reason for this is that it is serious limited in spindle speed when threading (typically if the spindle speed is above 200 RPM, it displays an error code on the LED display, and a "TOO FAST" warning on the video display (which is normally good for 16 lines of program and status).
    Good Luck,         DoN.
--
Email: < snipped-for-privacy@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| Voice (all times): (703) 938-4564

I cut some threads at 800 RPM, doesn't take long! The index comes around a lot more than a thread dial so it just looks like a continuous cut. About makes you sick after doing it by hand. My friend cut some hydraulic cyl. ends and got them down to about a half hour. I programmed it in the CNC lathe and it took 3 minutes :-)
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

    [ ... ]

    That is pretty good. How fast was the approach?

    Yes -- except that there are no dials on the Compact-5/CNC. Just small (8mm or so) ball screws driven by stepper motors with a short timing style belt (about 4mm wide) in a very tiny housing with no room to add any kind of index sensor. Once I modify one to use servo motors, I will have to re-do the couplings so I can add an index. I will really need a rotary encoder on the leadscrew to get position feedback anyway, and if there is one with an index encoder as well, that should work -- perhaps with the addition of a position sensor switch.
    [ ... ]

    That looks interesting, and may allow me to use EMC2 in a Sun UltraSPARC system running Solaris 10, instead of a PC running linux. If all else fails, I could install linux in the UltraSPARC system. The need for the ISA bus for the Servo To Go board was locking me to older PC hardware.
    Solaris 10 comes with built-in real-time capability, instead of having to jack up the linux kernel and run it on top of a separate real-time micro-kernel.

    Great!
    Hmm ... the Anilam Crusader II used entirely capacitive sensor switch areas on a rigid panel, IIRC. (It has been probably twenty years since I last saw one.) I wonder where the problems are in the front panel. (Or was the AUX button a real separate switch?)

    O.K.
    [ ... ]

    [ ... ]

    Even with the slower speeds for the Compact-5/CNC while threading, it is impressive enough threading to a shoulder.
    As for 800 RPM threads, I have cut those quite a few times on the purely manual 12x24" Clausing which I use. It is only when threading to a shoulder that I drop down into back gear for slower threading -- or really coarse threads.
    Enjoy,         DoN.
--
Email: < snipped-for-privacy@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| Voice (all times): (703) 938-4564

The RPM on the threads isn't that impressive but their seems to be no noticeable pause between passes. It goes in, takes off cutting the thread, backs out, rapids (200 IPM on the lathe) to start. It's as if your thread dial is ready 400 times a minute (my spindle turns 2 revs per 1 encoder rev.). On my cylinder rod end program it turns 1" stainless steel to 3/4" dia X 3" and threads 3/4-16 for 1-1/2" of the end, complete with chamfers and blend from threaded to flat portion, 1 manual tool change, and done in about 3 minutes. Seems that everything I have test cut on the lathe so far is done in about 3 minutes.
Another feature that could be of interest to you. You should be able to use EMC with stepper motors and encoder feedback. It's of limited because if a stepper looses steps, increasing the speed doesn't seem to help :-) What I like about my mill is the use of linear DRO scales for position feedback. When I first got the mill it had some backlash between the screw and table due to bearings needing shimmed. The combination of tach feedback to the drive and scale feedback to the control produced a stable and accurate system even with almost 0.1" of backlash (or course it wouldn't have been very good on climb milling!). For your lathes, it may be beneficial to at least let you know if a stepper skipped steps and fault out instead of perhaps crashing.
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
In rec.crafts.metalworking, you wrote:

    O.K. I thought that you had perhaps done that well on a single fast approach, which would have been *very* impressive.

    It is only a 5" swing -- and perhaps 10" between centers. There is a steel splash backplate which also includes the electronics, with the keypad and readout at the upper right-hand corner, and most other things high enough to be above the splashes.
    To zero it (for each tool) there ia a bed-mounted microscope, so you zero the Y and Z Axes for the first tool move to the intersection of the cross hairs, then swap in the other tools and write down the offsets as you move the tool to the same intersection. But there is no provision for storing the offsets in a single place in the program -- you have to enter the offsets *every* time you call up that tool. A royal pain.

    I've been collecting servos of various sizes, and have some which are not that much larger than the steppers, which stick out from the end of the bed and where the cross-feed crank would otherwise be, so there is room for *them* at least. No provisions for linear encoders, so I'll have to put a rotary encoder with each servo motor.
    [ ... ]

    Well ... that tells you which trace is likely open. Perhaps it is an open via, and simply filling it with fresh solder might make it work.

    O.K. When driving the knee -- is there automatic gibb locking between motions?
    [ ... ]

    Hmm ... what if you need to cut a 13-1/2 TPI thread? That might be difficult with the 2:1 coupling.
    Of course, an encoder directly on my Clausing's spindle would be difficult with clearance for a 5C lever-acting drawbar. So -- the trick there is to add another gear engaging the spindle's own gear -- and ideally a spring-loaded zero-backlash gear for that. Another use for the set of gear cutters and the index head on my (very manual) horizontal mill. :-)

    O.K. I could probably do that as fast on the Clausing, using the bed turret and a Geometric die head. 3/4-16 is within the range of the sizes which I have.

    I really don't want steppers at all. It was bad enough trying to cut a Morse 2 taper with the Compact-5. Not only was the resolution limited ot 0.001" or 0.01mm -- but it was measuring radial dimensions, so the diameter goes in steps of 0.002" - *that* I could see with my bare eyes. Servos, told to move at speed N for one axis and speed J for the other will cut a very straight taper.

    The Bridgeport BOSS-3 has two tapered roller bearings ground to fit together to mount the ball nut (which is rotated for the 'X' axis, with a fixed ball screw simply clamped at one end, so there is less whip with rapid moves). There is very little measurable backlash there.
    The 'Y' axis leadscrew does rotate -- but that is only for a 12" motion, not the 18" for the Y axis.
    The 'Z' zxis is a ground hollow ball screw around the quill, and another rotating ball nut to move it up and down, so the force is concentric with the spindle axis.

    Perhaps -- but I really want servos on those when I EMC them.
    Enjoy,         DoN.
--
Email: < snipped-for-privacy@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| Voice (all times): (703) 938-4564

I know what you mean, I prefer servos too. One thing I like about servos is that their torque is generally rated as continuous and goes up from the rating whereas stepper torque is rated as holding and goes down. I didn't know that the Emco lathe had ball screws, is sounds like it will be a very nice project for an EMC conversion. The hardware sounds great, just sounds like the control need updated a bit.
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

    [ ... ]

    O.K. So this is a hypothetical, instead of an actual bit of hardware. Yes -- I can see benefits to this -- when you need a lot longer Z travel -- perhaps to accommodate varying length tools in the quick-change spindle.

    [ ... ]

    O.K. Just a potentially longer wait to start the next thread pass. I kept thinking in times of gear synchronized leadscrews for whatever reason.

    Agreed. And another type of lathe which might be good for the pawn production is a tracer lathe.

    Remember -- I've got a fully working stepper driven Compact-5/CNC, and a totally spare machine with dead electronics for using to built the EMC-driven package. And I'm actually missing one of the steppers -- sold to someone else who was restoring another machine. So -- I can make this as a servo based lathe without losing the existing stepper based machine -- until I have everything working, then I swap over and look into making a second servo-based lathe so I can do things like second-operation CNC without having to reconfigure the first machine.
    [ ... ]

    [ ... ]

    Just as the hardware on the Bridgeport BOSS-3 mill is great for CNC (as it started life as a stepper CNC machine). The only problem is that the servo motors are longer than the steppers, so they won't fit where the steppers did for the Y-axis. Bridgeport made a recess in the knee to clear the stepper which pointed towards the knee's jackscrew from the Y-axis timing belt pulley below the Y-axis ballscrew. I need to make a new belt housing and motor mount, with the belt going diagonally to the right and down at about a 45 degree angle so the motor will stick back beside the knee, instead of into the cast-in recess. I've seen photos of newer Bridgeport CNC machines, with servos instead of steppers and with the diagonal belt to the servo mount. I just wish that one of those had been part of what I got with the machine.
    The original CPU on that machine is a LSI-11 -- the same 64K address space as the Compact-5/CNC -- but arranged as 32K 16-bit words, instead of 64K 8-bit bytes. (Later LSI-11 CPUs had memory management hardware, allowing splitting into 128 KBytes of address space, half for "text" (machine code) and half for "data" (the G-code and variables). But this one is too early a version for that.
    Enjoy,         DoN.
--
Email: < snipped-for-privacy@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| Voice (all times): (703) 938-4564

I always thought it would be interesting to have a small lathe I could put on my Bridgeport table and use a Servo and brake on the lathe spindle. Having a lathe combined with a 4th mill axis... I guess it isn't any better than a lathe with live tooling but probably a lot cheaper for a home shop.
RogerN
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
I agree that a big red 'OFF' button should kill everything, VFD included.
I tend to like all the controls in one spot with the panic controls on the top with the regular controls lower on the same panel. No sense having an operator have to think about if a situation is a big deal and have to move to a different panel. Milliseconds matter at this point.
Ignoramus25401 wrote:

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
My plans solidified a little bit. I will use the original START/STOP to supply or kill power coming _INTO_ the VFD.
I will install an Allen Bradley 800T style lever switch with XA switches, that has three positions, to provide FORWARD/OFF/REVERSE functionality. The lever switch has an extra convenience of being equivalent, on tactile level, to the direction of intended rotation (turn to the right for clockwise, turn to left for counterclockwise).
In addition I will mount some speed control pot near the lever.
i

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Ignoramus25401 wrote:

Could you take advantage of dynamic braking by turning off the power like that?
--Winston
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

No, but I think of that stop button as a safety stop to stop the machine when something goes haywire.
Under more normal circumstances, something else would be used to stop the spindle. As of now, I set it to stop in 0.5 second using dynamic braking. Any time shorter than that gives a weird feeling that the machine is fake.
i
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.