Milling Machine Design Request

[Ahmad_k]

Have you tried it with different motor, smaller one perhaps?

Are the mosfets from well known brand? Some imitation mosfet has very bad performance.

-ichan

If you looked to the schematic you will see that Mosfets are from IR. And for different motor, i also have 23H276-30-8A with same situation.

10KHz pulse that is very fast (10000 / 200 = 50 complete turn, for FULL STEP) and your lead screw is about 5 turns per inch than your system moves 10 Inch/Second.

You understood it in wrong way. My pulse clock are 500Hz but these saw tooth come from PWM signal when limiting the current.

1- Build simple PWM with a 555 timer which can vary frequency from very slow (zero if you can) to 10KHz (or higher if you wish).

2- Apply this clock pulse to your stepper motor driver.

3- Start with slow (motor should advance with each pulse).

4- Increase the frequency slowly until there is noise or missing step.

5- Your stepper motor can run up to the speed just before it skips step.

Here is a simple but popular RAMPING method that many peoples used:

Start RAMPING ---------------------------------------------------------- TOP SPEED                                   
__/’’’’’’’’’’\_______/’’’’’’’’\______/’’’’’’’\_____/’’’’’’\____/’’’’\___/’’’\__/’’\_..........

This is linear RAM up. Use it in reverse for RAM down.

You should start with FULL step first. If it works on full step, it will work on HALF step and so on (1/4, 1/8 …and more).

If the above testing method works then hook-up for MACH3 to see what is the problem. I think the CLOCK pulse frequency is the main problem.

And who said that L297 can do a microstepping ?

Guys, i don't want to repeat my situation in each reply. So please start reading from http://www.sonsivri.com/forum/index.php?topic=27112.msg101681#msg101681

The noise has nothing with Mach3 or the control board
the problem is with the driver board. The Biggest problem is the noise when enbaling the driver without applying clock This is my first priority.

[pickit2]

Have you only one earth point. you may have reflections on the moter leads, and your moter my stutter. have your FETS the protection diodes in the device if not you could try adding them.

[Ichan]

If you looked to the schematic you will see that Mosfets are from IR

An imitation of Rolex watch has Rolex written on it.

-ichan

Edit: Take a break, look around - you might visit this 20A Unipolar Stepper Driver site, with pcb layout and movie: http://www.dalton.ax/stepper/

[dennis78]

Did you try to add diodes between drains and Vm and maybe put additional didodes between drains and GND(fast diodes will do better)?
All experiments first try with short twisted wires between driver and motor.
Maybe this additional diodes can be more effective if you put on motor side connector.   
Some RC combinations between drains and GND can be good.
Try to increase C7 and increase resistance of divider for reference voltage.
Which RC filter do you add on sense inputs? Too small RC is ineffective, too much RC can  cause system unstable.

[Ahmad_k]

Did you try to add diodes between drains and Vm?
Some RC combinations between drains and GND can be good.

For diode, they are included in my first design. But the motor sunk 11A and there was no noise and no current limit.

For RC combination (Snubber) i was thinking about that, i will try it tomorrow. and i will try again the diode.

[dennis78]

You drive mosfets direcltly from HC outputs. I think 5V is very small voltage for properly driving power mosfets (i read specification and see 5V is enough,but...). Current capacity of 74HC08 is very small for properly driving power transistors. It can be big problem. Dynamical losses can be very increased. Did you try with some specific gate driver? Sometime I use simple PNP-NPN pairs for driving gates.   

   

[Ahmad_k]

You drive mosfets direcltly from HC outputs. I think 5V is very small voltage for properly driving power mosfets (i read specification and see 5V is enough,but...). Current capacity of 74HC08 is very small for properly driving power transistors. It can be big problem. Dynamical losses can be very increased. Did you try with some specific gate driver? Sometime I use simple PNP-NPN pairs for driving gates.   

   

This is a good point, I will see what to do about that, i will try NPN PNP driving method. But AND gate method is used in a wide application note and almost in every design i have found on the net.

[pickit2]

your Fets in post are Logic-Level Gate Drive, res on = 0.035 VGS = 4.0V, ID = 21A „
 but you could up the voltage on the 74HC08 and see if that helps, but I think the noise is due to ringing in you motors maybe an earth loop.

[dennis78]

...but you could up the voltage on the 74HC08 and see if that helps...

Increase voltage will not much help in dynamical process. Gate need much current in rise/fall time periods (sometime up to 2-3A, but very short time)  Increasing voltage can help more in R*I^2 losses, but i think he doesn't that problems.

... I think the noise is due to ringing in you motors maybe an earth loop.

And I'm very suspicious about grounding, because I had many problems with control and power grounds. Only connections control-power grounds in one point near IC's ground  solved some parts of problem. Sometimes just small parallel control wire and wire with high speed current/voltage changes made big troubles. This is part of my experiences in some of my previous projects with combinations analog/digital controls and high speed switching power elements/modules/devices.  Ahmad said that he good know grounding and other layout's problems and he was careful about it. I believe him.

[pickit2]

Ahmad_k you said without controler you had noise, have you put a scope on your fets and that aera to see whats going on there, a resistor snubber on motors may fix  reflected noise on cables, I remember spending a day on a scsi server that had two termanateors fitted. on one port and one only fitted when two was needed.

[Ahmad_k]

Ahmad_k you said without controler you had noise, have you put a scope on your fets and that aera to see whats going on there, a resistor snubber on motors may fix  reflected noise on cables, I remember spending a day on a scsi server that had two termanateors fitted. on one port and one only fitted when two was needed.

No, not without controller. L297 has three inputs. EN, CLK and CW/CCW. When Applying only the EN signal the motor start making noise.

Posted on: June 21, 2010, 10:48:09 22:48 - Automerged
This is my driver PCB image, almost all important net are named. There are a lot of modifications i made to this board. Like adding capacitor and removing snubber diode on the motor side. Adding filter to SENS pins.
What do you think about this design ? any problem that can occur ?

[Ichan]

What do you think about this design ?

Could be. My suggestions:

- Cut the ground track at the top which connecting ground from power section to the regulator
- Make a new ground track from the left connector directly to the center pin of regulator
- copper pour the ground on the power section (only)
- Never use white square power resistor as current sensing resistor, they are highly inductive i would prefer many 1206 in series
- Add 10 ohm smd resistor in series to the gate of the mosfet
- Make sure you use original mosfet  i'm serious.


-ichan

[oldvan]

Since having some of them oscillate and provide me hours of frustration tracking down the problem,
every 78XX regulator I use gets a pair of 0.1 uF 0805 ceramic capacitors directly at the device's leads,
one from input to ground and one from output to ground.  The 0805 size fits easily across the pads for
the regulator, and 5000 of the capacitors only cost me about $4.00 on eBay.



Placing a decoupling capacitor as close to the power pins of each IC as possible is generally good practice.
The entire 74** family of chips can be a son of a glitch to deal with when bypass caps are inadequate or
too far away.

On this board, U1 has (figuratively) Miles of GND trace, lengthy VCC trace, and decoupling capacitor
nowhere near.  U1 is being asked to quickly charge and discharge capacitive loads (the gates of the
MOSFETS) and needs thorough decoupling to provide it cleanly.  Try adding a 0.1 uF 50V ceramic
directly between pins 1 and 14 of U1.

[dennis78]

Oldvan is absolute right. I always put these capacitors just near IC's( all types digital and analog circuits because 1000 reasons). Sometime it isn't nice and professional, but...

I saw your layout. In this moment I don't have time for detail analysis. You should complete separate power/control grounds. I saw that some parts of your control system get ground over power ground. It isn't good. You can see my suggestions in attachment. Don't allow parallel and close wires(if you have that situations) with control and power signals. Never deal same wire for ground of both signals. In your design,you know better and easy way how to do it.
 

[Ahmad_k]

Could be. My suggestions:

- Cut the ground track at the top which connecting ground from power section to the regulator
- Make a new ground track from the left connector directly to the center pin of regulator
- copper pour the ground on the power section (only)
- Never use white square power resistor as current sensing resistor, they are highly inductive i would prefer many 1206 in series
- Add 10 ohm smd resistor in series to the gate of the mosfet
- Make sure you use original mosfet  i'm serious.


-ichan

I did all these stuff with no improvement at all.

And for capacitor i said that i have added some capacitor in the circuit for the regulator and for all IC.

I will take a break on this project for two days then i will comeback. I'm very tired from this driver

[oldvan]

And for capacitor i said that i have added some capacitor in the circuit for the regulator and for all IC.

- Add more decoupling caps : i have added more than 5 100nF and still noisy

Sure enough, you did and I entirely missed it.  Sorry about that. 

1) Have you looked at your 5V supply with a scope?  How quiet or noisy is VCC during operation?
    If no scope is available, a small speaker in series with a few uF of capacitor can give some clue for noise in the audio range.

2) Is the motor noise you are experiencing now worse in full step mode or in half step mode?

[Parmin]

Ahmad,

I would like to ask you to do something really stupid.  

Could you try and reverse the wiring of one of the motor coil?
ie, say flip the B0 and B1 coil connections.

I am sure you have installed the coil the right way, but if your motor emit sound as soon as it is enabled,
it occurs to me that the coils might be pulling the core in different direction (very minute distance) at high speed.

[Ahmad_k]

i have connected the motor windings as shown in the datasheet. but i will give it a try.

[abcsantosh]

Hi Ahmad_k ,

 one small solution from me. You said motor has noise when you work on INH lines. Try this,
The direction pin of L297 must be held active low or active high. You can't keep it just floating. Even when testing the motor. Connect it either, to gnd or to power line. Same is true for step pin also. This means your breakout board will must have this feature, instead only providing isolated hi and lo signals. Use well designed breakout board.
 I am using this board for my machine. http://pminmo.com/4axisopto/4axisDIYopto.htm

 Once I had this kind of problem, When I was checking my first motor controller with L297+L298. I had left the direction pin open and gave the step pulses on step pin, and motor kept running noisy. But when I connected dir pin to gnd or to power, motor runs smoothly.

 Also you have to configure proper voltage and current limit. Otherwise it won't work. The circuit consisting of L297+L298 gives best low cost solution to beginners.

 Have a good day

[Ahmad_k]

Hi Ahmad_k ,

 one small solution from me. You said motor has noise when you work on INH lines. Try this,
The direction pin of L297 must be held active low or active high. You can't keep it just floating. Even when testing the motor. Connect it either, to gnd or to power line. Same is true for step pin also. This means your breakout board will must have this feature, instead only providing isolated hi and lo signals. Use well designed breakout board.

If you have see my schematic you will find the answer.

[abcsantosh]

  Hi Ahmad_k, next time I will read and watch carefully. I wanted to help you.

  Now just try this. Open Mach3 > Config Menu > Motor Tuning. In the box Step Pulse 1-5 uS, increase the pulse to 10 and see what happens.
  You are using 74HC14. It is Hex inverting Schmitt trigger. May be the default pulse width from Mach3, which is too narrow, is not enough to trigger it. Increasing the step pulse width may be helpful.  So just try above mentioned solution.
  
  Have a good day

[Ahmad_k]

  Hi Ahmad_k, next time I will read and watch carefully. I wanted to help you.

  Now just try this. Open Mach3 > Config Menu > Motor Tuning. In the box Step Pulse 1-5 uS, increase the pulse to 10 and see what happens.
  You are using 74HC14. It is Hex inverting Schmitt trigger. May be the default pulse width from Mach3, which is too narrow, is not enough to trigger it. Increasing the step pulse width may be helpful.  So just try above mentioned solution.
 
  Have a good day

Again Please read.

[sadman]

Hi Ahmed

Please check out this

http://www.4shared.com/file/vqI_4Cvj/CNC_MACHINE.html

regards
sadman

[Parmin]

Ahmad, don't keep us in suspense! 


Have you find the problem? what is it? can you solve it?

[Ahmad_k]

Ahmad, don't keep us in suspense! 


Have you find the problem? what is it? can you solve it?

Ohhhh No, still have the buggy noise but the machine works perfect. Just be patient until i upload all data.