My CNC questions thread - keep an eye on this one...

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Dinkum

Well-Known Member
Joined
Mar 16, 2012
Messages
63
Reaction score
4
I am trying to build a CNC dividing head, but I have absolutely no knowledge of the electronics side of things that doing this will require. So I have decided to start a thread just for asking various questions regarding CNC basics rather than cluttering up the Q & A forum.

So here's my first dumb question:

My understanding of CNC basics is this: You have a breakout board which creates the pulse, width, and direction signals and these you send these to the motor driver. Then the driver converts these into the necessary format for driving a stepper. The driver also can act like a relay and feed more power into the motor than is contained in the p/w/d signal. The driver also can do fancy stuff like microstepping. Is this all correct so far?

Here's a second question: what sort of controlling input is this driver looking for? ( http://www.ebay.com.au/itm/Dual-H-B...Electrical_Test_Equipment&hash=item2ebd9899bb ) It does not take a p/w/d signal, does it? Do I need an L297 in the circuit somewhere?


Lastly, where is a good tutorial for this sort of thing, please? I mean the overall circuit of a stepper system rather than just which wire goes to the gecko board type of thing.


 
hi Dinkum,

found this about BOB http://www.cncroutersource.com/breakout-boards.html should answer more than a few of your questions.

its the PC that will send pulses to the driver/controller (gecko) through the BOB then the driver/controller (gecko) sends the steps to the motors.

yes gecko's can micro step.

i have never heard of a driver acting like a relay and boosting more power into the stepper. the power for the steppers come from the power supply. the driver just send the pulses to the steppers to tell them what to do..... i.e step and direct.

the controlling input comes from a program like Mach 3. there are others and there may be a simple one just to run a single stepper.

don't really know what you mean by your last question. do you mean how it all goes all together? if that's the case then......

you need controller software for your pc...... mach 3
mach 3 takes the g-code and turns it into pulses.
the pulse then goes out of your pc to the BOB.
the BOB sends it to the Gecko's.
the Gecko's then do there stuff and sends the pulses to the steppers.
your steppers move.


hope that helps..... this cnc stuff can be a right mine field when you first get into it.
 
Thanks.

I can understand all of this except for the "gecko" part. Is this just a driver/controller? What cheap alternative is there to it?

Is it possible I can use the driver board I get with a stepper motor taken from a copier?








 
Blighty has it correct. Mach generates the pulses, the BOB is really simply there as isolation to prevent overloading the PC parallel port, and help keep higher power noise from feeding back into the PC.

There is another element that you ought to consider. The early versions of Mach generated the pulses and output them through a parallel port. There are issues of timing and accurate control, and PCs with parallel ports are getting hard to find. So a few 'motion control' boards have been developed. These boards do the fast 'real time' stuff. The most common example in the Mach world is the SmoothStepper. These devices communicate to the PC via USB or ethernet, and eliminate the need for a BOB.

Think of the stepper drier as a power amplifier, like in a stereo. It receives the low power pulses form the BOB or the motion controller and it supplies the high current to run the motor. A driver must be matched to the motor- if you use a 3 amp motor you need a driver that can supply 3 amp. Gecko is one brand of driver, and one that is very well respected and of moderate cost.

L297/L298 are real old, real wimpy drivers. There are MUCH better drivers around now, and the prices are not bad. A driver form an old printer or copier will also be wimpy- think about the work they are doing, and understand whoever engineered it took the smallest, cheapest motor that would do the job. Its unlikely that motor/driver will be much good on anything bigger than a CNC pencil :D

The world of CNC is very broad. You need drivers and that can get you into chips and power transistors, etc, and it runs all the way up to some VERY complicated software. You need to pick your battles carefully. With drivers relatively cheap I would not fool with circuits and chips- buy a driver and get on with it. Often I read about someone wanting to build his own boards, and write his own software- my guess he will never get a working machine.

When I make metal things I dont feel the need to smelt my own iron- I buy that stock and concentrate on making parts from it. When I do CNC I buy modules and hang them together, let someone else make circuit boards.

As to a nice tutorial all in one book I am afraid there is not one. There really ought to be, and I have thought about writing one, but it would be a BIG job, and the field would probably change a couple times before it got written. Dig around the net, this forum has some CNC stuff, but there are entire forums devoted to just CNC. Its a lot of work and digging, but the info is out there.

Good luck!
 
Dinkum said:
Thanks.

I can understand all of this except for the "gecko" part. Is this just a driver/controller? What cheap alternative is there to it?

Is it possible I can use the driver board I get with a stepper motor taken from a copier?

sorry about that, in you first post you seemed to be using driver/controller/gecko as a description for one thing. yes there are a lot cheaper ones out there, but as your going to use it for a dividing head you need to pay attention to micro stepping and gearing. stepper motors will move 1.8deg per step. so you need 200 steps for one rev or 360deg. so lets say you have 1 to 1 gearing, this would mean you couldn't have any movement the you can't divide 1.8deg into. with 10 micro steps this now changes to 0.18deg and so on.

you could probably use the driver board in the printer, but chances are that part of the circuit will be incorporated with other bits that run the printer. I'm sure some one who knows a bit about electronics can help you out, but as RonGinger says, its best to buy with things like this. as Ron also said about people making everything DIY. i know some one who has done this and the end result was not so good. after all, they have been designing and upgrading Geckos for 10 odd years. what makes him think he can do the same thing in 10 minutes.
 
but as your going to use it for a dividing head you need to pay attention to micro stepping and gearing. stepper motors will move 1.8deg per step. so you need 200 steps for one rev or 360deg. so lets say you have 1 to 1 gearing, this would mean you couldn't have any movement the you can't divide 1.8deg into. with 10 micro steps this now changes to 0.18deg and so on.

Microstepping is NOT a way to get more resolution. The micro steps are not as accurate as you need. Microstepping is very important to reduce resonance issues when motors start moving faster. It make motors smoother, not higher resolution.

A dividing head should use gear ratios to get the step resolution they need.
 
RonGinger said:
Microstepping is NOT a way to get more resolution. The micro steps are not as accurate as you need. Microstepping is very important to reduce resonance issues when motors start moving faster. It make motors smoother, not higher resolution.

A dividing head should use gear ratios to get the step resolution they need.

if thats the case, then how come my mill wont move 0.01mm with out micro stepping. the only way to get my mill to move 0.01mm with out micro is to gear the motor 10 to 1.


i have a 5mm pitch screw, 1 to 1. one step at 1.8deg will move the table 0.025mm. with micro stepping @10 micro steps. i now get 0.0025mm. table also moves up to 7500mm/min full rapid.
 
Your first ebay item, the one in the black box, looks like one of the Leadshine drives. These are well respected drives. I have used several from the US supplier keling.

The small one looks to be a single chip version of the TB6550 chip. There are about half a zillion of these built in China, to mixed reviews.

In particular the 3 or 4 axis version have some real issues around the opt-isoloater. The ones on the board are way to slow causing missed step problems. It is also bad to use a board like this with a BOB that also has optos on it- you do not want two optos in series. There is quite a good analysis of this board on the SeeMeCNC yahoo group including directions for modifying the board to jumper out the optos.

MY personal opinion on these boards is that they are actractive because they are cheap, but they are cheap, and in the long run Id spend a few dollars more up front and get a better product.

I would not even remotely consider trying to modify the internals of a drive to get higher performance. If it was practical the manufacturer would have done it.

Nema23 or Nema34 refers to a motor mechanical spec, it says nothing about the current required for the motor. You first pick a motor big enough to supply the mechanical power you need, then find a driver that can supply whatever current that motor requires. There are Nema34 motors that take less current than some 23's.
 
RonGinger said:
Microstepping is NOT a way to get more resolution.


This is very important and never mentioned. You can use micro stepping to increase resolution, but you must understand the drawbacks. The biggest one is the reduction in torque while micro stepping. Mircostepping Myths

Steps Torque
1 100.00%
2 70.71%
4 38.27%
8 19.51%
16 9.80%

While you can micro step to increase resolution, you're giving up torque. This means you are either required to use very large motors, or position feedback to make sure you aren't loosing steps under load. Large steppers are the answer as lacking torque you might never get to position. Huge steppers really seem great on the spec sheet, but the biggest result of that is wasted power.

It would seem 5 phase steppers would be much better with 2.5 times as many steps as a 2 phase stepper, but the drives aren't common.

The torque issue doesn't become apparent unless you calculate axis force and work back to motor torque, knowing the reduction in torque due to microstepping.

Greg
 
dieselpilot said:
This is very important and never mentioned. You can use micro stepping to increase resolution, but you must understand the drawbacks. The biggest one is the reduction in torque while micro stepping. Mircostepping Myths

Steps Torque
1 100.00%
2 70.71%
4 38.27%
8 19.51%
16 9.80%

While you can micro step to increase resolution, you're giving up torque. This means you are either required to use very large motors, or position feedback to make sure you aren't loosing steps under load. Large steppers are the answer as lacking torque you might never get to position. Huge steppers really seem great on the spec sheet, but the biggest result of that is wasted power.

It would seem 5 phase steppers would be much better with 2.5 times as many steps as a 2 phase stepper, but the drives aren't common.

The torque issue doesn't become apparent unless you calculate axis force and work back to motor torque, knowing the reduction in torque due to microstepping.

Greg

i remember reading something along those lines.. think it was on the Gecko website. it was sort of a happy medium between micro stepping or not, because of the points Dieselpilot made. if this is correct, then that would mean my 960oz steppers are only putting out 140oz ish... that cant be right. ???

it was basically saying that if you ran a stepper on full steps you will get 100% holding torque, but 70% running torque. when micro stepping (10) you will get 80% holding torque and also 80% when running (as i said can't remember the exact figures, but its something like that.) should really find the web site........ back in a bit ;)
 
That is a very interesting article about microstepping. Its summary says
In summary, although Microstepping gives the designer more resolution, improved accuracy is not realized.

That strikes me as a real dumb statement- you get more resolution, but no improved accuracy? So you can say to have 256 micro steps but they wont be where you think they are?

I stand on my view, microstepping does not give useful resolution, just bragging rights.
 
Hey Blighty

blighty said:
leaving all the techno stuff out.

if you want your mill to move 10mm and it does. it can't be to bad can it?

Yup...provided it really moves....if you lose steps and not know it...it's not helping....then an encoder would at least help to let you know it didn't move..... ::) or the control can compensate ....maybe....

Dave

 
then an encoder would at least help to let you know it didn't move

now that might open up a huge can of worms :eek:

as far as i know, it don't work like that. the pc tells the motor to move 2000 steps, then the encoder says it actually moved 1995.....i don't think the pc
would then sends another 5 steps to make all ok.

reason for this is.....

i used to work with pick and place machines. if at any point it mist a step it would call up a "drag error" and the machine would stop. you would then send it home (so it knew where it was) and restart the program. it would then work for another 10 minutes before something else would make it stop ::)...... so if it was just a case of sending another 5 steps, then the machine wouldn't of stop.
 
Your recollection is the same as mine Blighty....and getting an error like that in the middle of a part would not be good.....

Dave

Actually I never did like steppers!.... ::)
 
blighty said:
i used to work with pick and place machines. if at any point it mist a step it would call up a "drag error" and the machine would stop. you would then send it home (so it knew where it was) and restart the program. it would then work for another 10 minutes before something else would make it stop ::)...... so if it was just a case of sending another 5 steps, then the machine wouldn't of stop.

This sounds almost exactly like how us union folk work at Boeing... except we don't get sent home =/

We have a problem, then we write an NCR (non-conformance report).. move on to something else until the engineers get back to us... find some other problem we don't feel like working around... write an NCR...

- Ryan
 
blighty said:
if you want your mill to move 10mm and it does. it can't be to bad can it?

It depends on how accurate you want it to be. You might be fine with 10mm. I'm OK with 10.00±.01mm, and I think my machine is actually pretty close to this on a finish pass, maybe even a touch better as the axis repeatability is 2 or 4 microns, but I'm only setup to resolve 2 microns. Th biggest problem for may machine is rigidity.

Your 140oz stepper output is plenty. This is 225 pounds of axis force. If the cutter hasn't made the cut it probably won't before the next move is made. The problems arise when you have cutter load and part of the torque available is overcoming friction and inertia to move the axis. Properly size steppers do work well, but sizing steppers at full step torque then using microstepping to boost resolution would be a mistake.

I ran through all these figures when I got my Emco CNC lathe. It had 120Ncm 5 phase steppers (500 steps per rev) and actually resolves 2.5um using half stepping and making 2kN of axis force. Emco claimed 1micron resolution, but the machine obviously doesn't have it. I looked at 2 phase steppers and figured I needed 280Ncm steppers at 5 microsteps to obtain the same resolution and axis force. I decided on 100Ncm servos with 5000 line encoders for 1um resolution and 25% more axis force. Will I really be better off? I don't know, but I know the servos are quiet, and my servos have never missed a step.

For a worm or harmonic drive reduction dividing head where the ratio is substantial, I think microstepping should work OK, but you probably wouldn't even need it with that much reduction.

As far as the PC knowing whether or not it made it to the commanded position, it depends on how it's set up. The stuff we work with in hobby CNC doesn't do that, there is no feedback, even with servos. Gecko servo drives will send an error if the servo lags more than a given amount. Stepper drivers though are completely open loop. If it doesn't move, the PC doesn't know and carries on with the program. Your pick and place machines had some sort of feedback and could detect an error. This can be done with EMC2, encoders on the steppers, and the proper motion control cards. EMC2 with the proper hardware supports full feedback motion control where the motion control card closes the loop in the servo rather than something like a Gecko servo drive. Full closed loop takes a lot more effort, and think I'm going to try it on the Emco lathe.

Greg
 

Latest posts

Back
Top