Need help with jog speed in Mach3 motor tuning !!!
 

Hi, I am need of some assistance from the group mind.

I have got my machine running around but not cutting yet.
I went into the shop today to tune the motors and set steps per movement.
Set the steps as per the step spreadsheet on this forum and got pefect measurements as far as I can tell so far.
I then set the speed and accel , I am in mm motors ran and sounded good as high as 9000-10000 and accell about 500.

The problem is when I save my settings and go to jog or run one of the sample files, the machine is not responding .
Meaning it is leaping about , motors running rough.
When I jog at half speed it is rough sounding 100% jogging gives same results as running a file.

Shouldn,t at least the 100% jog speed be as smooth as the motor tuning setup.

Any assistance much appreciated.

Andrew

what motion control software are you using?
Read the manual & everything will be there.

As in the question , am using mach3 .
Nothing mentioned in the manual about discrepency between settings in motor tuning and actual running / jogging of the machine.
It probably is a simple software issue , but for the life of me I cannot find a reference to a similar issue.


My first port of call was the manual , was just hoping that someone has had experience with a similar issue.

Andrew

It might be computer issue, had one IBM Intel 3.0, it was working ok but when running had strange "clicks" during motion.
Also some had problems with BOB or LPT cable resistance, search the forum I cant find exact thread.

Thanks for the suggestion , but not sure about it being the computer.
I should have been a bit more detailed.
Upon first getting the machine moving without tuning the motors for correct steps etc, I was able to run code / jog no problem. But only fairly slowly.
Upon tuning and setting the correct steps am I having issues.
The machine moves smoothly and sound great under motor tuning, but on exiting that dialog the machine doesn't move the same at all.

Andrew

The values you have chosen are quite conservative at 10000 and 500 so it is unlikely that is your cause.
Note that Mach3 needs to be restarted before the motor tuning takes effect.
You might also think about increasing the kernel speed to 35000hz in Config, Ports and Pins if you are using a reasonable speed processor in your machine controller PC - a restart is also required here.
You might also wish to run the drivertest.exe file in the root directory of the Mach3 install to test that is working OK.
Double check the gear speed calculator and the steps you have entered in Mach3 on all axis including the slaved axis on the X.
Also double check the velocity and acceleration values are the same or at least what what you intended for each axis.
Note you must press the "Save Axis Settings" button for the values to stick.
Lastly sometimes a delay in the Step Pulse of 2us and 5 us in the Direction Pulse might settle things - these settings are also located in the motor tuning dialog in Mach3.

It's a start.

Regards
Ross

Thanks Ross

Have tried all your suggestions, no change.
Did the kernel speed test . Get a perfect as you can flatline at 25000hz or 35000hz.
I even bumped up the speed to see what would happen, under motor tuning I can run at almost 13000 on the Y axis and still sounds feels good.
Same deal, when I exit the motor tuning dialog ( making sure to save changes ) restarting, it wont jog or run code smoothly. I can jog at low % but even then sounds rough.

The readouts are corresponding to a correct amount of movement in mm, but its almost like its trying to get there too fast. I am wondering if its ( mach3 ) is running in inches somewhere and I am getting some sort of speed multiplier . Clutching at straws, but that is what it looks like. The motors are trying to move it too fast, that is what it looks like.

Any more suggestions would be much appreciated as this is frustrating to say the least.

Andrew

Just to add a bit more info.
Oriental motors 7.2 gearbox 36 tooth pinion module 1 rack. 203v geckos
Steps per mm set to 127.324.
speed 9000 accell 500 ( fairly conservative ).
Have tried jogging a single axis and that is still a no go.
Tried 25000hz and 35000hz. no difference

Thanks Andrew

Rival,

You mentioned in your first post that you went into the shop to tune the motors. What does that mean to you?

Does that mean you changed the pot setting on the Gecko drives so that the motors actually sound or feel smooth?

If you are having a problem at a higher motor tuning value, slow things down and see if that makes a different.

Just a suggestion.

What Mach3 version are you running? Maybe its some test version?
Do you have any plugins installed?
Pendant?

At the pulling hair out frustration stage now !
Have tried downloading and installing a fresh latest mach3 . No change.
Gone over my earth wiring, changed printer cable.
Check wiring connections.
No plugins installed
No pendant.

Have released the springs so motors run free.
They seem to turn then lock up under jogging, releasing the key and starting again they turn then lock up.

But under the tuning dialog box they turn smooth as silk.

Any wild ideas ?

Running out of things to try

Andrew

Did you clean your PC registry? Sometime the unknown junk in the PC can give endless hassles...

Hi Andrew
I do not know what I am talking about but are all the motors locking up if not maybe could be a faulty driver or wiring to driver, also I think it pays to have a computer just dedicated to mach 3 no other software

Regards Neil

Cleaned the registry , no change to situation .
Computer is dedicated to machine , no internet connection.
Only other software is a trial version of aspire.
I dont think it is the drivers as the motors run fine under the tuning dialog as mentioned.
If I run code and slow the feedrate to a crawl it runs smoother, but I cannot imagine being any good at cutting much.
If I reduce the jog speed to 30% or so it moves around but still not as smooth as under the motor tuning section.

Anyone else got even the remotest sugestions.

Thanks Andrew

Have you tried the Mach support forums?

Hi Andrew
There is a fair bit of probably fairly relative information on artsoft site ie troubleshooting video. Here is the link
http://www.machsupport.com/videos/

Regards Neil

Thanks, have made a post on there . See what they can come up with.

Hi Andrew

Unfortunately I have no knowledge of the Gecko tuning process you are using.
Makes that part a big blind spot for me.
Logically though, it is possible that the two processes eg tuning and machine movement are being driven by two different software applications.
Under that circumstance it is possible for one to run like a dream and the other to just refuse to work.

You have not yet provided specific information re ports and pins in Mach3, it is pretty common to screen shot or record these settings for others to sort through for you when problems occur.
Go to Ports and Pins, then record the details under the 3 tabs - Motor Outputs, Input Signals and Output Signals. The other three tabs generally work ok as per the Mach3 installation defaults.

To speed things up...
1. run and test on one axis only, preferably the Y axis with the single motor rather than look at the machine as a whole.
2. start disconnecting things like the VFD
3. reduce complexity where ever possible.

You will smile all the wider once it all works, sometimes walking away and fighting another day brings a fresh mind to the problem.

Regards
Ross

Can you post some pics of your control box? Also how are the ends of your cable runs teminated?

More info
 

Hi, Here is a couple of pics of control box.
Also a vid showing difference between tuning running and jogging

http://i982.photobucket.com/albums/a...o/100_3888.jpg

http://i982.photobucket.com/albums/a...o/100_3889.jpg

http://i982.photobucket.com/albums/a...o/100_3890.jpg

http://i982.photobucket.com/albums/a...o/100_3891.jpg

http://s982.photobucket.com/albums/a...chjogissue.mp4


I have grounded all the shielding back to one point . All the connections to bob and geckos are crimped bootlace ferrules.
Check power at all points . geckos getting good 56v from power supply.
Bob getting good 5v from power supply.

The only odd thing ( may not be odd ) should the bob have any led that light up when the box supply is off and the printer cable is connected. I presumed this just indicated that was getting power.
But I thought I read another post about the bob getting power from the pc was bad.

Hope this all helps as got me stumped at present.

Had a few responses from the mach3 forum but nothing that has solved the problem.

One guy asked what the "time in int" was in the diagnostic page. I posted him the numbers, what should they be ?

The stray black wire is the limit switches which are not connected yet.
I have run grounds from the table/gantry/car back to common point.

Thanks Andrew

Enjoy the video !

Andrew
Did you manage to look at the trouleshooting video (see post 16 for link)

Regards Neil

By the sound of motor locking up I would only think of some mechanical error or low voltage PSU. It would not lock up like that by mach not working correctly, don't know why the difference in tuning dialogue and in program directly, but you should measure you PSU voltage while motor turning (pulling current from it) to see can it keep up with them. Also review the current settings on gecko that may have the same effect.
Motors stall like that usually when they are under powered (electronics or psu error) or too weakfor the task (mechanical).

p.s. I missed the nice toroidal transformer in the pictures, check the gecko resistors (for 3Amps its 35.25k or Orange-green-orange and it does not look like that in the pics), and try to eliminate that clicking mechanical sound.

Danilo ,

Thanks for the thinking.
I checked the power supply while running multiple axis, no drop in voltage.
The resistors are .25w 47k 5% tolerance, which should be fine.
I dont think the motors are lacking power. If I run code and crank the feedrate down so it doesn,t move to crazy, I can lean all my weight on the gantry and has little effect/no effect on movement.

Andrew

I'm out of ideas for now.

I have not followed all the threads in detail, but could it not be a similar problem that I had with Brad's machine.
Are you using all 8 wires from the stepper?
If you are using only 4 wires, are the other wires properly screened and isolated ?
Could it not be a problem that if these other motor wires are just touching each other while the gantry is moving this could act as a breaking system ?? Just a tought

Kobus, Thanks

Just went and triple checked, no stray wires touching .
Only 4 of six wires connected ( checked a gazillion times )

Andrew

Suggestions from the mach3 forum , seem to indicate it is most likely a computer issue. Likely the computer, which has an intergrated video card is not giving a good pulse stream.
Thinking is that under the tuning dialog the screen does not get refreshed, hence the motors run smoother. Once out of the dialog the video uses up too much cpu time.
That is the thinking, now to hunt down another computer ! not an easy task.

Andrew

Very reasonable answer, you can try maybe without new computer, there is a command to turn off toolpath display (that is usually what consumes graphic or cpu time)
Ger21 has it in Mach 2010 screenset. Try to find that command it may help diagnose faster.

Are the wires shielded for step and direction? (I see red wires going to ground for the motors) I would also unbundle the step/direction and motor drive wires until you are sure you are not getting cross talk. You may even want to backup a few steps and pull your motors and use short cables from the Gecko drives and do a "Kitchen Table" test at this point.

Give us a pic of how your motors are wired.

I do not think there is anything wrong with wiring or power supplies......

Mach runs in two modes:
- motor tuning, via pulse streams generated by the PC and then it runs good and strong
- actual cutting, also via pulse streams generated by the PC and then it runs bad

So the fault lies in how the PC makes those pulse streams for the 2 modes of Mach

All the cables to and from the geckos are shielded cables.
The red wires are the ground terminations for the shields for the cables from the machine to the geckos.
The input signal cable shields are also grounded to the same point.
The wires from the pmdx to the geckos are shielded cables but I have not ground those.
I think the explanation from the mach3 forum seems the most likely.
I did individually test each motor / gecko today by detaching all the relevant wiring to each motor / gecko.
No change and all motors performed the same fashion.

Motors are wired bipolar.
Checked the connections any number of times.

Thanks Andrew

Thanks Gerald , I am pretty sure ( hoping like hell ) that the pc is the cause of my woes.
Going to try and beg borrow or buy another buy the end of the week.
Watch this space.

If any one else has any suggestions though , would appreciate the input.

Need to get this puppy up and going as have at least 2 paying jobs lined up !

Andrew

Given that the entire setup works fine in tuning, but doesn't in normal mode, I agree with the assessments that it is a software / computer issue.

Note that if there is an expansion slot (PCI, etc) in the computer, you can still add a video card even if there is onboard video. The BIOS will have an option as to which card it defaults to. So if access to a video card is easier than a new computer, you could try that.

Can you tell us anything about the computer you are using? processor, motherboard, memory?

Brad

As below for specs on computer, was thinking about adding a video card.
But only paid under a 100 aussie dollars for the tower. going to cost me almost that for a video card. Might see if I can get another tower with a seperate video card. Then I can swap the video card to give me a third configuration.

Existing computer as below
Make/Model
HP/DX2000
Processor/Speed
Pentium 4/3.0GHz
Motherboard/Chipset HP/865G Chipset
Memory slots
2 DIMM sockets
1024mb/2Gb maximum
PC3200, 400MHz, DDR1 SDRAM memory supported
Installed memory 1024mb
Storage 40gb Hard Disk Drive
Optical devices CDRW
Modem None
Operating system Windows XP Pro (no media supplied)
Video Intel Extreme Graphics 2 (integrated with the 865GV chipset)
Sound
Integrated Intel® Audio with Premier Internal Speaker
Network
Integrated Broadcom NetXtreme Gigabit Ethernet NIC
On board I/O
6 USB ports (two in front, four in back), USB 2.0
One 9-pin 16550 serial ports
One 25-pin parallel port port, support EPP and ECP operation
Keyboard, monitor, and mouse ports
Microphone input, line-out, line in audio jacks
Other
PCI 2.3 Three (3) full-height slots on PCA
Expansion bays
Two (2) 3.50" half-height
Two (2) 5.25" half-height
Physical
355 x 180 x 410 mm
Weight fully loaded - approximately 10Kgs
Case condition Some minor marks
Warranty 30 Day


I think the logic of the movement in tuning proving the hardware / wiring is sound = the computer at fault

Andrew

Hi Andrew

I have read the mach support posts here for anyone who is following.
My reading of this is that it is hardly conclusive, however at this stage just like PC repair it is time just start swapping out bits (including trialing another PC).
Even a laptop might be the go for a quick test if it has a LPT port.
You have not mentioned an operating system clean install as part of your diagnostics, sometimes this can be useful.
Ensure the drivers esp the onboard video drivers are correctly loaded.

Read your PC spec and it is considerably more powerful than what many others are using to run Mach.
If I was to say one thing it is a little short on memory particularly as it has shared memory for the onboard video.
The fact remains however, many machine controllers use onboard graphics quite successfully - including mine!
So a quick swap out with any PC you can get your hands on, then install mach still might be the quickest way to see where you stand.

Regards
Ross

Try this: Go into the BIOS and disable the audio, and the network card.
In XP, set the video to the lowest resolution you can stand (Think recovery mode).

Note that when looking for a replacement video card you're not looking for the latest fastest; the oldest, basic PCI slot simple video card is desirable. If it doesn't have a fan, it's perfect. The kind of thing a PC enthusiast put in a scrap pile after an upgrade; street value around $10. Anybody around you repair PCs?

Ross / Brad

Thanks for following the thread and giving a fresh look at things.
The Computer was bought as a reconditioned pc.
Was a very basic install with just window xp.

Hope it is the computer, as otherwise i'll be pulling wires !


Thanks Andrew

One thing to consider if XP was already loaded you do not know what was loaded with it. They usally use univeral type images that may have drivers for many devices installed that could conflict somewhere. A fresh load would be a good test also, even on the second PC your buying. Also as with a cheap stand alone video board you can also buy a cheap stand alone LPT Port.

Note that the issue here isn't the overall processor speed; it's the latency of the interrupt response of the machine, which is affected much more by the peripherals and memory bus than the processor. The issue with machines with integrated video, and in particular machines with certain generations of Intel video, is that the video accesses to main memory starve the CPU. So the very fast CPU is ready to context switch and respond to the timer interrupt, and suddenly it can't access main memory to do so and has to wait. I believe the 865 platforms are one of those.

Other peripherals can also cause this effect, which is why you troubleshoot by disabling as many of them as you can. Also, any power management settings should generally be turned off, except you may want ACPI in general enabled, even though all the subsettings are off. Reducing the display resolution, and in particular the number of colors, reduces the amount of bandwidth video is using, and might get you past the latency problem, although adding a display card is the solution you need for this symptom.

Good luck locating a PC to comparison test with.