Hello Guest it is August 03, 2020, 08:07:54 PM

Author Topic: Accurate homing  (Read 20216 times)

0 Members and 1 Guest are viewing this topic.

Re: Accurate homing
« Reply #20 on: August 21, 2007, 01:02:59 PM »
Yes, my limits are junk. No question there. E-stop halts the pulse stream cold and hard. No question there.

Does Mach have the capacity to do all I'd like it to do? That is the question.

To that end,
Does anyone know if the encoder DRO's are consistent across error/reset events?
(So far the consensus is "yes," they are consistent. Does anyone know for sure?)

Ps, And philosophically, Mach is not about getting what you pay for (a cheap controller). It's about lowering the barrier of entry for quality machine control.

-Ted
« Last Edit: August 21, 2007, 01:12:38 PM by Erichtg »

vmax549

*
Re: Accurate homing
« Reply #21 on: August 21, 2007, 01:27:28 PM »
OK so WHY do you use the ESTOP to stop the machine???  WHat you are saying doesn't make a whole lot of sense, please explain a little futher maybe we are missing something here. You can EASILY set mach up to do a controlled stop with a switch. You can call the switch anything you like even WHOA if you want.

Do your limit switches stop the machine when you hit them? Obviously they do or you would not be complaining about them(;-) They are not home switches they are limit switches. If you need accurate home switches install accurate home switches??????

Everything you have asked about is possible but you still refuse to adknowledge it????????

Ya might want to reread the part about building the "machine and program code" to be as dependable and reliable and accurate as you WANT the software to be(;-)

(;-) Are you sure you are not my youngest son hiding behind that screen name(;-)

(;-) TP

« Last Edit: August 21, 2007, 01:32:58 PM by vmax549 »

Offline Chaoticone

*
  • *
  •  5,628 5,628
  • Precision Chaos
    • View Profile
Re: Accurate homing
« Reply #22 on: August 21, 2007, 02:12:00 PM »
Quote
Does anyone know if the encoder DRO's are consistent across error/reset events?

Ted,
    I will find out the definitive answer ASAP and let you know.

Brett
;D If you could see the things I have in my head, you would be laughing too. ;D

My guard dog is not what you need to worry about!

Offline jimpinder

*
  •  1,232 1,232
  • Wakefield, West Yorks, UK
    • View Profile
Re: Accurate homing
« Reply #23 on: August 21, 2007, 02:19:30 PM »
I  wasn't infering that Mach3 was a cheap controller - it has all the features of the professional systems I have seen - Fanuc being the leader. What I was saying is that the machines I have seen are way in advance of mine. They have feedback, proper limit switch systems etc etc. I saw one the other day with two chucks, that could pass the work from chuck to chuck. I haven't got that because I don't want to pay for it.

As far as I can see, the DRO's have no feedback (certainly using steppers). Logically they must therefore take their position from the pulses put out to the motor (try running the program without the miller connected - the DRO's still register) - they know how many pusles move the axis how far. BUT if the computer puts out a pulse which the motor misses - for what ever reason, then the DRO IS NO LONGER ACCURATE. As an alternative - if you press E-Stop - and the computer instantly ceases putting out pulses - yet the momentum of your system carries the motor over a few steps (4.8 to the thousanth of an inch on mine) then the DRO (which registers to 1/10th of a thousanth) can no longer be accurate. I suppose in theory if you belt the E-Stop at one end of travel, then belt the E-Stop at the other, the two could cancel each other out.

If you set up a key on your keyboard to stop travel - (I was going to suggest the space bar but this might not be feasable) - then that is quicker and easier to get to the trying to click a mouse on a screen icon. That would take care of E-Stop and cut out a lot of problems. If you set up your soft limits that should take up another problem (because you shouldn't bang your limit switches anyway).

The other question about homing - have you set up your homing as per the tutorials - and seen if it works. You can use your limit switches - because when homing the limit switch function is cut off - until the homing is complete - and then it resumes.

I think you have to do all you can before adding other bits and pieces to try to fix a problem that may not be there.
To my mind , keep it simple.
Not me driving the engine - I'm better looking.
Re: Accurate homing
« Reply #24 on: August 21, 2007, 02:34:13 PM »
Quote
Does anyone know if the encoder DRO's are consistent across error/reset events?

Ted,
    I will find out the definitive answer ASAP and let you know.

Brett

Thanks Brett,

For the record, I'm curious about both encoder feedback DROs as well as Machine XYZ DROs - - Does Mach keep track of exactly how many pulses were sent or received across error/reset events?

-Ted
 
« Last Edit: August 21, 2007, 02:36:01 PM by Erichtg »

Offline Chaoticone

*
  • *
  •  5,628 5,628
  • Precision Chaos
    • View Profile
Re: Accurate homing
« Reply #25 on: August 21, 2007, 02:38:03 PM »
I understand what you are trying to do, will let you know ASAP.

Brett
;D If you could see the things I have in my head, you would be laughing too. ;D

My guard dog is not what you need to worry about!

Offline Chaoticone

*
  • *
  •  5,628 5,628
  • Precision Chaos
    • View Profile
Re: Accurate homing
« Reply #26 on: August 21, 2007, 03:01:05 PM »
Yup. Mach always knows and tracks the number of pulses. The only reason position is ever lost is if the motor itself cannot stop that fast. Another way to test what you are trying to accomplish I think,  hook up an encoder, assign it a DRO, place in e-stop and see if the encoder dro updates as you turn it. I don't think you need to test it though. That answer came from one of the immortals.  ;D

Brett
;D If you could see the things I have in my head, you would be laughing too. ;D

My guard dog is not what you need to worry about!

Offline ART

*
  • *
  •  1,700 1,700
  • Tough as soggy paper.
    • View Profile
Re: Accurate homing
« Reply #27 on: August 21, 2007, 03:03:32 PM »
Hi:

  Just to kick in.. ( Thanks brett fro steering me to this post..)

  Estop and limit events shoudl not be prevalent on any system, BUT, if for whatever reason they are..

Estop is just that, an emergnecy, all power shoudl be cut, ( which woudl release brakes on a bug system, thats truly safe..). Mach3 keeps track of all pulses at all times. The DRO's are always accurate as to the pulses put out. Doesnt matter what you do. Postion is lost only when a motor cant keep up. If using servo's, the Rogers baord makes it pretty much infalible. Brian has demonstrated at shows with grabbing the axis and stopping it, thus causing an estop, when the rest is pressed, the DRO's go to the actual motor position, and no homing in necessary. That having been said, Homing shoudl never be risky, it shoudl be accurate. Mine is accurate to .01mm's. If yours isnt, then you need better switches.

   Now it IS possible to home to an index using macro's. All it takes is a "move to index " macro. A very slow move that stops when an input sees a pulse woudl do. Slow, but do-able, Ive never bothered as my homing always seems so accurate I woudlnt gain from the complexity. KISS is always a good idea if you can do it IMO.

  When hitting stop, many steppers will lose position due to abrupt stopping. Many servo's will not. Alot depends on the system.

>>they know how many pusles move the axis how far. BUT if the computer puts out a pulse which the motor misses - for what ever reason, then the DRO IS NO LONGER ACCURATE. As an alternative - if you press E-Stop - and the computer instantly ceases putting out pulses - yet the momentum of your system carries the motor over a few steps (4.8 to the thousanth of an inch on mine) then the DRO (which registers to 1/10th of a thousanth) can no longer be accurat

  Both true statements. BUT , and this is where the theory breaks, it is just as likely for a system to miss a step pulse as an encoder pulse. Both are similar in nature, and both happen. Since nothing, ( truly nothing ) can see the missing encoder pulse, no system, closed or open loop, servo or stepper, is immune to lost pulses. Whiel its true that in true closed loop only 1 weak link exists, any chain is only as strong as its weakest link, so closed loop is questionably better. The true ( and to my mind only )  advantage to a closed loop system with fuill encoder feedback, is that Estops, and sudden stop button presses will not lose position. That having been said, that situation shoudl be a rare one. And homing will cure that in a properly setup system.  I may make some changes soon in the stop system, but Estop will always be a problem.. and should be. :-)

Art
Re: Accurate homing
« Reply #28 on: August 21, 2007, 06:25:50 PM »
WOOOOOHOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!!!!!
Thanks so much Art and Brett!

That's really great news. That really opens up the possibilities.

I'm upgrading my BP to AC drives and will look into writing a macro for finding the index pulses on the encoder lines for super accurate homing.
Probably something like:
 - fast "ref all zeros" to limit switches
 - custom jog to get close to the index
 - slow find index.

I probably won't end up using Rogersmachine encoder board because it does the same thing as the error signal from the servo drives. I can wire the encoders to an additional parallel port myself.

But to be sure. Steppers + encoders + Rogersmachine encoder board is a great solution I wish I knew about a long time ago.

Art, you should advertise the DRO functionality (input and output) in the documentation. I have friends who are slow to adopt Mach because they aren't sure of what is going on under the hood.

Yeah I've been abusing the e-stop for a while. What function was attached to the space bar in the standard screen for Mach2? I don't trust it for some reason -- can't remember why exactly, but I do know for sure there were a few occasions when hitting the space bar didn't have the desired stop/pause effect to make me go for the ESC key by default.

For everyone worrying about the safety of my high speed system: It is unsafe, in that it has no breaks. However it's a bench top system, so it can't kill anyone, AND it has an enclosure that must be locked for motion to occur.

Thanks everybody!

Ted

Ps While I've got Art's attention... Is it possible to enter numbers in Mach 3 using the keyboard's number keys? I haven't figure out how and it's my biggest obstacle to upgrading from Mach 2. I've only a greasy touch-pad attached to the controller now (but would prefer keyboard only).

« Last Edit: August 21, 2007, 06:29:27 PM by Erichtg »

Offline Chaoticone

*
  • *
  •  5,628 5,628
  • Precision Chaos
    • View Profile
Re: Accurate homing
« Reply #29 on: August 21, 2007, 06:37:36 PM »
Np Ted,
   
Quote
Is it possible to enter numbers in Mach 3 using the keyboard's number keys?

Do you mean in a dro? Hmmmmmmmmmm not sure if you don't have a mouse. I just click in the DRO, type the number and hit enter. I can't remember trying it any other way.  ???

Brett
;D If you could see the things I have in my head, you would be laughing too. ;D

My guard dog is not what you need to worry about!