Cutting mdf doors, laser coming? #1 - Sri Lanka

Gerald, In a earlier mail I could remember you asking me as to what model of motor I am using and also the way that I have wired it. You aked me these questions in order to advise me on what value of resstor to use on the driver.

The motor I use is PK296A1A-SG3.6 and I have two pairs of wires going into the two coils A and B.

TOday is a day off for us. But I would like to gather as much info and advice possible so that things would be easier for me tomorrow to see as to where I have gone wrong.

To adjust the midband resonance you have set motor speed to 1/4 rev per second and adjust the pot until you hear a distinct null in the motors vibration. Am I correct?

That motor is specified as 1.0 Amp when wired bipolar or 1.5 Amp when wired unipolar.

It has 6 leadwires - only 4 of the six wires are ever connected:

- if yellow and white are not connected then you have it as bipolar 1.0 Amp

- if other colours are not connected then it is unipolar 1.5 Amp (In this mode you only use half of the windings)

Set the resistor according your choice of wiring (it doesn't make a huge difference for a CNC router). Get the wiring wrong and you kill the gecko...

Gerald, Do you have to provide a good earth to the table as well. At the moment it is not earthed and its sitting on a concrete floor.

Yes, that is how to adjust the driver/motor resonance. Remember you have gearbox on the stepper - the 1/4 rev/sec is for the motor only.

The ground/earth issue is illustrated here. Use the earth point in the control box

Gerald, As you assumed I too feel that this so called erratic behavior of the motors is due to electrical noise.

One final question the earth point (what you have reffered to in your earleir posting) in the control box is the alu plate onto which the Gecko drivers are mounted. And also the negative wire of the power supply (connected to the transformer) should also be connected to the same alu plate. The earthing leads comming from the table (x,y,z) should be connected to the same alu plate.

Please correct me if I am wrong.

Yes, I use that plate for all the grounds/earths. Also the cable shields connect to the plate.

The table x,y,z earth wires are not separately coming to the plate - they are connected in series. If you start from the alu plate, a thick copper wire (+6mm²) goes to part 1010307, from there to 1020456, from there to 1030422 and finally from there to gauge plate in the z-slide

Thanks Gerald.

Well done Fabrica,

credits goes to your courage, promptness and hardwork which braught you a success. though I am having all the facilities yet I could not comple the work. I had got all the material today and I will be starting the table very soon and start the new thread within few days, as I got inspired from the Great Mr.Gerald and you I hope I will be the next one to complete the "MechMate Beast" Soon.
..Manjeet

Manjeet Happy to hear that you are making some progress.

Gerald, After doing everything that I promised to to do today I still have this erratic behavoir on the motors.

I gounded all wires running from the Gecko's to the motors to alu plates. Connected the negative of the power supply to the same alu board. X,Y and Z were grounder in serise to the same alu board. Tuned the pot on the Gecko's to the minimum humm levels(to overcome midband resonance).

The symptoms on the final run I made in the evening were as follows.

While jogging (with only one axis working) the steppers were running without any problem on all three axis. When you run the machine on a programme where all three axis start to work together we get some motors to miss their steps and direction specially at the accelataration stage of moving.

We took the router wiring completly out of the control box. Even the on/off switch was disabled and was done manually at the router point. Still nothing changed. We closed shop at 7.00 pm today. Tomorrow out first job would be to take the transformer out of the box and see weather this component is the culprit of undue noise generation. At this point we are almost sure that what we are encountering are noise related problems.

Even after taking the transformer out if the problem still persists I would replace the PMDX card to see weather this unit is diffective.

If the the acceleration factor in Mach3 (motor tuning) is set too high, then it is completely normal for a motor to get overloaded and lose steps during the accel. stage of the movement. But then it will do it also when runnning only one axis....

Does your power supply perhaps lose voltage when it has to push 4 motors? Is there a regulator in your power supply that starts to clip/oscillate? The transformer is maybe too small? (should be around 300VA - looks much smaller in the photo). You say it is a standard power supply in your company - can you give each axis its own supply as a test?

Fabrica,
Is your computer fast enough? An underpowered computer would cause these same symptoms.

DocTanner

Very true Doc. Thanks. (How is your animal behaving itself?)

At this point I don't think it is an interference issue.

Extremely unlikely that the PMDX has anything to do with it....it is actually a pretty dumb device. MUCH more likely the PC driving the PMDX.

My settings for acceleration for x and Y axis is 500 mm/sec and for Z it is given as 1500 mm/sec.

The transformer which we are using was just lying around in our factory. It has a wattage of around 1500 Watts. This is putting out 48 V AC into the PMDX 135-8020 power preparation module. The powder comes out of the prower prep module after getting multiplied by a factor of 1.66.

If everything fails I will provide each axis a individual power supply.

The computer is a Dual core running at 2.8 Ghz with 1 GB RAM.

Could any software on the PC be interfearing with the signals.

Fabrica,
Your computer is more than adequate. There are a lot of things that interfere with Mach3. It wants fulls control There is a file on Mach's website http://www.machsupport.com/optimisation%20XP.txt Backup your system first!!!

Doc, Up from 4.00 am trying to figure out this problem.

Since you pointed me towards the problems that I may have on the computer side. Below given are my observations while setting up mach 3.

The mach version I use is Version R2.0.015. Does this have a special driver for PMDX cards. I could not find the PMDX option in the drop down list while configuring the mach 3. On the device manager list I found the PMDX card listed as a pulse generator.

Does the PMDX card have a driver for it. It is definitly not plug and play.

Gerald, As you very well know the motors that I use are Oriental PK296A1A-SG3.6. On the motor it mentions a stepping rate of 0.5 degrees per pulse. Is this after considering the Gearing ratio.

I understand that normal steppers operate at 1.8 deg/pulse.

Microstepping is done at the Gecko drives is it?

It is very early here...have to got to work. The PMDX card has zero intelligience, zero memory - it cannot even talk to a PC to tell that is plug&play. A PC could never detect its ID. It cannot use a driver (too stupid), so there is no driver for it. You could connect the printer cable direct to the geckos (some people do this) but the PMDX makes the wiring easier - that is why it is called a "parallel breakout card".

The 0.5 degrees is after considering the gear ratio.

Gerald, In the morning cleaned up the computer as per the intructions given in the link provided by doc.

We feel that their is slight improvement in the smoothness of the motors.

Now when engaging the router on a samlla cutting programme (without tool) the motors are moving smoothly when cutting straight lines. When cutting arcs all three motors on x and Y axis stagger and move.

Gerald, on the motor tuning and setup screen you get two inputs called step puls (1-5) and dir pulse. What is the normal velocity you use for all axis.

Cutting speeds are mostly in the range 1800 to 3600 mm/min. Sign letters in MDF typically 2400 mm/min. Just cutting a lot of straight shallow grooves at 7200 mm/min.

Gerald now the only problem (for the moment) that I am facing is that while only one axis is moving (cutting on a straight line) the movement is ok and smooth. When two or more axis are moving (cutting arcs and circles)the whole gantry is jerking.

Gerald, on the motor tuning and setup screen you get two inputs called step puls (1-5) and dir pulse. Whet do these two inputs mean.

I have just started to move the bulky transformer and place it outside the control box.

On Saturday afternoon I asked if the problem was speed related and you said it was not....

It is known on ShopBots, and also on the MechMate, that the gantries and cars start to resonate at higher speeds when cutting arcs and circles. This problem is definitely speed related.

From the mechanical side it is critical that the V-rollers sit very correctly and firmly on the rails. The Y-car must have the spring roller (M130100) on the side away from the motor.

But, the above problem does not cause the motors to lose/gain steps. You are telling me that you are gaining/losing steps?? We have to be very clear about this.

If you cut around a full circle, and you have the erratic problem, does the cutter finish at the starting point, or is there an error?

error = electrical problem
no error = mechanical problem

"on the motor tuning and setup screen you get two inputs called step puls (1-5) and dir pulse. Whet do these two inputs mean." I think it is a way of lengthening the pulses coming from the PC - have never used them so I don't know.

On saturday afternoon Gerald we were working on the jogging mode. We ran the machine fast and slow but the problems that we encountered happened while running fast as well as slow. That is why I informed you that speeds are not linked to the problems encountered.

We checked the V rollers and also the bearing fitted to the opposite side of motor mount on Y car. They were ok.

To get a conformation on weather this was mechanical or electrical related problem we disengaged the Pinions from the racks and ran a circle cutting programme. All three motors are jerking while turning. We are now very sure that this is a electically related problem.

We also cut a full circle with motors engaged to racks although the movenet of the gantry is Jerky the circle was ok it ended up at the same place.

We are still working.

Trouble Shooting:
Disengage motors by removing the springs
Gantry should go from one end to the other with a firm push
Y-car should do the same
Z-axis should remain stationary with motor spring removed.
All v-wheels should remain in full contact during all of this. Movement should be very smooth.
With springs replaced, Move the gantry. It should take a firm push to move.
Reverse movement of the gantry. There should be no "slack" when going from one direction to the other.
Slack indicated a loose pinion gear
Do the same for each axis.
Use pin connector for every connection to breakout board and Gecko
If you don't have a quality crimping tool for pin connectors, then it's best to solder then cover with heat shrink tubes.
Motor wiring: From Gecko to motor:
Pin 1 Ground
Pin 2 Power <---possible culprit "I'm using 70 volts from toroidal transformer"
Pin 3 Coil A = PK296A1A black
Pin 4 Coil A~ = PK296A1A green yellow Do Not Connect- Insulate from any contact
Pin 5 Coil B = PK296A1A red white Do Not Connect- Insulate from any contact
Pin 6 Coil B~ = PK296A1A blue
Pin 7 Not Connected
The PMDW-122 board is recommended, but a simple breakout board will work for trouble shooting
Pin 8 Direction - From Breakout board
Pin 9 Step Pulse - From Breakout board
Pin 10 "Common" 5 volts <---- from computer or breakout board (correctly jumpered PMDX-122)
Pin 11 resistor 8.2K 1/4 watt resistor
PIN 12 other end of resistor

If everything checks out, then go to www.machsupport.com/videos.htm They have some great tutorials.

Hope it helps.

"we disengaged the Pinions from the racks and ran a circle cutting programme. All three motors are jerking while turning. We are now very sure that this is a electically related problem.
" Well done! That is a very helpful diagnosis. The problem sure is electrical or software.

Because the circle completes fully and there are no errors in position, I believe that interference is not the problem.

It looks like Mach3 and your PC are having problems giving evenly and correctly spaced pulses to the steppers.

I hope that you have the Mach kernel speed set to 25kHz. Config > Ports & Pins. I have heard that you shouldn't set it higher unless you have a very good reason.

May I suggest that you take this problem to the Mach support forums?

Gerald,

Problem sorted out!

I do know that in the process of sorting out my problems I blew your fuse on sevaral occasions.

My computer guy whom I recruited to the design jobs on Artcam had accidently enabled the Auto limit overide button on the settings page (press Alt6 on main menu). He has done this to overcome a problem at some point of our testing stage and forgoton to switch it back to the disabled stage.

I feel that when we are overiding limits on mach 3 the programme goes on to a safe mode and due to this does not obey many of the commands that go through a G code based sequence of events (programme). This is why the machine behaved well while running on the Joging mode and not on a programme mode.

Just to prove a point I thought of uploading the above given pic. These cuts were made before the exact prob was identified and also at the time of these cuts the motors were very gerky in their movements. Can the pictures prove any of these misbehavoir.