Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Ricky 2019 on May 01, 2019, 09:34:34 PM
-
I am very new at this, I have 34 stepper motors and on the Y & A axis I get stalling /growling when in the jog mode and also when trying to run the demo in mach 3. Mostly when I stop and start. I have check settings, wiring, looks right. My x and z seems to do find. I have changed settings in motor tuning and I have to go back to my original setting. Can anyone help?
-
I am very new at this, I have 34 stepper motors and on the Y & A axis I get stalling /growling when in the jog mode and also when trying to run the demo in mach 3. Mostly when I stop and start. I have check settings, wiring, looks right. My x and z seems to do find. I have changed settings in motor tuning and I have to go back to my original setting. Can anyone help?
You have FIRST to make sure that the microstep settings at the Stepper drivers are set correctly first.
Check that it is set to 1600 steps per rev. (NOT ANY SETTINGS HIGHER than that)
-
Check motor tuning. It seems tat your acceleration is too high.
-
when you say micro step settings at the stepper drivers are set correctly first, you mean the switches? if so i have done that and did change them and had to go back with the original settings. but not sure how to set the 1600 steps per rev. can you explain how?
ill also check the motor tuning again.
-
when you say micro step settings at the stepper drivers are set correctly first, you mean the switches? if so i have done that and did change them and had to go back with the original settings. but not sure how to set the 1600 steps per rev. can you explain how?
ill also check the motor tuning again.
Yes the switches.
What is the model of your driver?
For example, if the driver is DM860T then its SW5 SW6 SW7 SW8 should be ON OFF ON ON
(see attached)
YOUR driver should have it's own chart.
-
my drivers are DM860A
-
my drivers are DM860A
See Attached
-
as i said i am new at this so i may ask questions that i guess i should know but, what about 1 through 4. i don't see it.
i did have the steps per rev. higher. I do thank you for your help.
-
as i said i am new at this so i may ask questions that i guess i should know but, what about 1 through 4. i don't see it.
i did have the steps per rev. higher. I do thank you for your help.
Please see attached to set SW1 SW2 SW3.
I am guessing that your motors are rated 3.5A max current.
Are your motors running better now?
-
Yes they are 3.5 A.
I have not be able to get back to the shop yet to reset things as of yet, hoping to today, for sure tomorrow.
My settings was my problem i guess, i was told to set my steps per to 4000 so i'm hoping that was my problem.
I'll post my results when change them.
Thanks
-
Made changes on driver switches but did not solve the growling/stalling. May be I just need to find tune the motors. Any suggestion are welcome?
-
Made changes on driver switches but did not solve the growling/stalling. May be I just need to find tune the motors. Any suggestion are welcome?
see attached (mm units) as starting point. Then tune up by making changes to suit your hardware.
If you have a fast PC, step pulse could be decreased to 5 or even 2. Decrease until SMOOTH.
increase speed and accel for each axis until it's becomes too high then back off 20% for normal use.
-
should my drivers never be set above 1600 steps per rev. ?? And than set accordingly?
-
Hi,
it is possible to set your microstepping regime higher than 8 microsteps per full step (1600 steps per rev)
but you get little advantage.
It is seductive to believe that you get increased resolution as the number of microsteps go up, unfortunately
that does not happen in practice. In fact no genuine increase in resolution is achieved beyond half stepping (400
steps per rev). If no increase in resolution then why bother? The answer is smoothness of motion....the higher the
microstepping the smoother the motion.
If however you increase the microstepping too high then the required pulse rate to spin your motors will exceed what your
controller is capable of.
For example if you set your microstepping to 64 microsteps per fullstep (12800 steps per rev) and you want a maximum
speed of 600rpm from your stepper the required pulse rate is:
600 / 60 =10 (revs/sec)
Pulse rate =10 x 12800
=128kHz
Machs parallel port in default is only 25kHz and there are only a few external controllers that can generate pulse stream at that
speed. If however you choose 8 microsteps per fullstep the pulse rate is:
600 / 60 =10 (revs/sec)
Pulse rate = 10 X 1600
=16kHz
ie well within Machs parallel port even at default speed.
As it turns out the smoothness of an 8 microstep regime is very nearly the same as a 16 or 32 microstep regime
ie there is little to be gained beyond 8 microsteps (1600 steps/rev). As a consequence 4, 5, 8 and 10 microsteps
per fullstep are the norm among Mach CNCers.
Craig
-
should my drivers never be set above 1600 steps per rev. ?? And than set accordingly?
Craig already said it well.
See attached and ask yourself.
Do I want to lose more "holding torque" by setting micro-steps much higher?
Holding torque.
A stepper motor's holding torque is the amount of torque needed in order to move the motor one full step when the windings are energized but the rotor is stationary
More holding torque is needed if you want to cut MORE deeper in one pass.
-
Also see attached (result published at Hackaday)
Highest accuracy is 1/8 microsteps (1600 microsteps/rev)
Above 1/16 microsteps (3200 microsteps/rev) stepper motor can no longer position accurately.
-
in motor tuning I have set my/ steps per, to 1600/ and I have changed my velocity and acceleration to many different numbers. I can get the motors to run smooth when I jog, but when running the demo in mach 3 they sound terrible. At this point I am not worry about speed. I just can't get them tuned. I know i'm just missing something but can't figure it out. I don't give up very easy so i'll keep at it. So any information on tuning is appreciated. you need to dummy it down, as I said I am very new. Thanks for all your information.
-
in motor tuning I have set my/ steps per, to 1600/ and I have changed my velocity and acceleration to many different numbers. I can get the motors to run smooth when I jog, but when running the demo in mach 3 they sound terrible. At this point I am not worry about speed. I just can't get them tuned. I know i'm just missing something but can't figure it out. I don't give up very easy so i'll keep at it. So any information on tuning is appreciated. you need to dummy it down, as I said I am very new. Thanks for all your information.
steps/per are not set to steps per motorrevolution it is steps/unit (mm or Inch, dont know in witch Units you are working).
do you realy have 1600 steps per mm or Inch?
are you realy sure about your Motor Connections (A+/A- and B+/B-)?
witch exact Motors are you using?
-
Hi,
I can get the motors to run smooth when I jog, but when running the demo in mach 3 they sound terrible.
First lets dispel a myth, Mach3 Demo and Licensed Mach3 run identically. The only limitation that need concern you is that Demo
is limited to 500 lines of Gcode at a time. The kernel frequency of Mach3 Demo is limited to 25Hz, but this need not concern you,
unless, as previously explained, you choose to set a too higher microstepping regime; if you keep to 8 microsteps per full step
or below the limited kernel speed will have no effect.
The fact that the motors run smoothly when jogging tells us that the steppers and drivers are good, it also tells us that the BoB
and Machs parallel port can produce the required high quality pulse stream. The fact that the same steppers/drivers/BoB run poorly
when processing code suggests to me that Mach's parallel port is running poorly when the Gcode interpreter is commanding
motion.
My guess is that when the Gcode interpreter is running it introduces a heavy demand on the PC's interrupt system which in
turn causes excessive jitter within Mach's parallel port driver, itself an interrupt driven timer.
Proving that this is so is not straight forward.
Would you please post a screen shot of DriverTest.exe. Its not conclusive but is still a fair indication as to the health or
otherwise of the parallel port driver. Note that some PC's seem to do a good job with a parallel port and another machine,
possibly with much higher specification, does poorly. There is no way to tell short of installing Mach and trying it.
Have you considered an external motion controller such as an Ethernet SmoothStepper or a UC300? One of the
advantages that an external motion controller offers (vs the parallel port) is very high quality, high speed pulse
streams WITHOUT being dependent on PC hardware. Should you consider this option you should budget somewhere
in the $150-$300 range, if you buy cheap Chinese rubbish you'll have even less luck with it than the parallel port!!!
Craig
-
Perhaps speed (in motor tuning) is set too high so G0 moves in program run are stalling /growling
Jog may appear Ok because the MACH3 Factory default Jog speed is at 50% (unless it had been changed by user).
-
Hi,
Jog may appear Ok because the MACH3 Factory default Jog speed is at 50% (unless it had been changed by user).
True....however this from OP:
in motor tuning I have set my/ steps per, to 1600/ and I have changed my velocity and acceleration to many different numbers
rather suggests that he has tried a range of accel/speed that might have given acceptable results.
Craig
-
without any preceise data (Motor Installation, Settings Motortuning, Speed of used G-code ...),
this is all a walk in the dark IMHO.
-
I am using inches and using longs 34 motors and I have checked connections several times. I did change to 50 percent, however it does appear to be faster when running the demo but it could be my imagination. I bought this as a kit and at time didn't know as much about it and as you see still don't know. Maybe it is the computer telling it what to do or maybe a differnent board would be the answer. I understand that without looking at the system its hard to diagnose. I am not giving up just have to rethink, reset, check and try again.
Thanks for everyone's help.
-
you can do quick checks by using the mdi Input line before run gcode
set your x axis Zero
then G1X10 F10
and G1X0 F10 for going back
the you can increase the F value and see when the stalling starts
-
ok i'll give that a try.
-
I went into MDI and added the gcodes, it stall very first thing, had to increase the F to 13 but was very slow and still not smooth. that was the best setting I could get. So went back to motor tuning and again one at a time I could get to run pretty smooth. but when added other motors, they started stalling and growling. No progress again today. Could it possible that the breakout board is bad??
-
I went into MDI and added the gcodes, it stall very first thing, had to increase the F to 13 but was very slow and still not smooth. that was the best setting I could get. So went back to motor tuning and again one at a time I could get to run pretty smooth. but when added other motors, they started stalling and growling. No progress again today. Could it possible that the breakout board is bad??
1. Please printscreen your Motor tuning page for x-axis. Paste into Paint save as .png format and upload picture here.
2. What Power supply do you have? Please let use know model.
Thanks
-
here is the motor tuning and power supply.
-
power supply
-
@ricky,
1. Are you using a USB or Ethernet external motion controller? If not, you cannot set step pulse and dir pulse to 0 and 0. Try 5 and 5.
2. If your are using metric units the Steps per unit (mm) should likely be around 320steps/mm Try 320 first then run axis calibration (settings tab) to auto correct later when linear motion is getting smooth. If using inches it will be much higher.
3. Set speed to about 1800mm/minute and accel. about 400 as starting point.
3. Your PSU is 350W. is only enough for two stepper motor at the most.
Best regards
Reuel
-
I will give it a try. I four power supplies, one for each driver.
-
O yes I am using inches.
-
O yes I am using inches.
Then try steps per inch 8128 first.
velocity about 60
acceleration about 12
After axis can move, you need to run axis calibration to confirm whether 8128 is correct or not.
-
and how are rhe Motors connected to the axis ? belt driven ? ballscrew ? any gearings ?
-
just ball screw. And I am using the printer parallel port. I will try these things you mention this afternoon. Thanks again.
-
just ball screw. And I am using the printer parallel port. I will try these things you mention this afternoon. Thanks again.
ball screw direct connected ? witch pitch ?
-
They are direct connected and the pitch i have for gotten, would this tell you (Bo6x92YPT8) if not i'll have to try to find that out.
-
goggle knows nothing about Bo6x92YPT8
-
This is what i bought: Shzond RM1605-1500 Ballscrew BF12/BK12, but i looked and i can't find your question out. Maybe i don't know where to look. Pitch is the number of threads per inch right?? If so is measuring with tape measure accurate enough to get answer??
-
so you have buyed am metric ballscrew (RM1605 means 16mm Diameter 5mm pitch),
but want to use imperial Units.
so we have to do some calcs.
we microstepping 1600 steps per motorrevolution -> 5mm
so steps/per (motortuning) is: 25.4/5*1600 = 8128
for a test use:
velocity 60
acceleration 12
and make sure Config ->Select Native Unit is set to Inches (if you have to Change, restart Mach3)
@reuelt, do you have a Magic glas ball ? ;D
-
After more research i realize i was incorrect in my statement. As i said i am new at this. thanks again and i will give it a try.
-
Hi
please post a screenshot of DriverTest.exe.
Craig
-
After setting motor tuning to your recommendations fine tuning (X axis) , its the best its done but i can run the mach 3 demo maybe 3 times without any problems but the next 3 times it will stall or have what sounds like vibration. it does it at no certain place in program. Check the volts while running nothing drops. When motor is running for several seconds in program without changing directions , the motor sounds like it is changing rpms, not a lot but just enough to hear. I tried the Y and A axis without any success. After trying Y and A and going back to the X axis had to fine tune again.
-
I am having trouble loading the pic of the driver test still working on that.
-
Ricky,
What is your PC spec. Is it a laptop or a desktop?
What is your CPU model and size of RAM and Video Card.
If you have a good CPU on desktop, try to increase the Kernel Speed to 35,000 or even higher.
My Win 7 can run at 100,000 with no problem because I have a Quad Core I5 7600 running at 3.5GHz. I can even reduce the step and direction to 2 and 2 and run at 4,000mm/min. With my older PC, core-2 duo running at 3Ghz, I can only run smoothly at 35,000 kernel speed but my pulse and dir usec had to be set at 10 and 5 otherwise motor will stall when G0 is executed. Also max speed had to be reduced to 1800mm/min
With my 2GHz laptop, I can only run at max speed of 1200mm/min (about 47inches per minute)
Time is money so a GOOD CPU will save you a lot of money and headaches.
-
it is a desktop, don't know model and size, not looking at right now, but will try to increase the kernel speed. thanks, headache is right.
-
Hi,
the main data to be retrieved by DriverTest.exe is the timing jitter of the interrupt timer. On most
PC's on which I have run Mach's parallel port the jitter is about 4us.
If you have Mach's kernel speed set to 25kHz the period is 40us. The jitter is therefore 10% of the period.
If you set the kernel speed to 35kHz with the same 4us jitter the percentage is 14%, ie worse.
The Mach wisdom passed down is that 'run Mach at the lowest kernel speed consistent with the minimum required
to meet the maximum output pulse train determined by the motor tuning'. Mach's parallel port runs more smoothly at lower
kernel speeds not higher.
Craig
-
here is the driver test screen.
-
Hi,
that is a very good plot. The average jitter is 3.2 us which is very good.
I would expect this PC to run smoothly.
Craig
-
so does that indicate that my problem would be in the software or maybe the breakout board? just trying to narrow it down.
-
Hi,
good question.
The fact that the machine does perform well on most occasions but not all suggests to me that the hardware is
OK.
I am still of the opinion that Mach3. or rather the parallel port driver, is going cranky at certain times. The DriverTest.exe
is very good indeed but what it does not do is show how the parallel port performs when 'under load' ie running Gcode.
I would now be trying to observe the step output pulse with an oscilloscope. That may not be easy. Even modest variation
can cause what looks to be a very poor trace. There are some scopes that collect histogram data about the signal under
test but they are rare beasts but wuld be ideal for this purpose.
Craig
-
Hi
One reason I advise to increase the Kernel speed is based on INFO from Mach3Mysteries.pdf by Art Fennerty.
"The MACH3 planner controls the CV'ing of the lines, as well as taking care of making each output equal to 5 times the kernel interrupt time (kernel speed)."
So MACH3's "planner" has to run slower to match any slower kernel speed.
Art even said
"I don’t mind admitting that a 60Khz run even with motors that don’t move
beyond 20khz is the smoothest option to use. "
BUT (faster kernel speed) has to be traded off to processor speed (Computer CPU speed) and the power required to fill the buffer faster than the engine can empty the buffer.
So, I get a very fast CPU that can give me "EXCELLENT" drivertest result at 100,000Hz then INTENTIONALLY slow down the kernel to 60,000Hz (instead of 100K) and MACH3 will run very well.
To run at 25,000Hz kernel means MACH3 CV planner has to be slow and limited to 200ms.
So MACH3 will run more smoothly in CV mode if kernel speed is higher than 25K (provided you have a fast enough CPU to avoid any buffer under-runs).
-
Hi,
who cares about CV, at the moment OP is struggling to get the motors to run under certain circumstances
and I don't think kernel speed is the answer. Vast numbers of users run Mach's parallel port at 25kHz without
problem....so why can't OP?
Craig
-
Hi,
who cares about CV, at the moment OP is struggling to get the motors to run under certain circumstances
and I don't think kernel speed is the answer. Vast numbers of users run Mach's parallel port at 25kHz without
problem....so why can't OP?
Craig
I read his reply #42.
I thought he is already running better now but sometimes having some CV look ahead jerks which can be solved by increasing Kernel speed.
-
Hi,
I read the same post:
its the best its done but i can run the mach 3 demo maybe 3 times without any problems but the next 3 times it will stall or have what sounds like vibration
That does not sound like a CV issue to me. Sometimes it runs and then on other occasions 'it sounds like a vibration'.
I guess that 'sounds like a vibration' is excessive jitter in the pulse stream.
Proving it is hard and solving it harder still.
Personally fooling around with a parallel port is so last century....get an ESS and never look back!
Craig
-
I think we are nearly getting there.
Originally his problem was because he had set the pulse & Step width to 0 u secs.
Attached is Electrical Characteristics of his driver which is based on Toshiba TB6600HG.
It says min. pulse width is 2.2 u secs.
(HIS DRIVER IS NOT LIKE A GECKO DRIVER which only need a Negative edge trigger)
So I have told him to set to 5 and 5 (assuming there will be some PCI bus latency loss too).
Jogging should now be OK (subject to a bit more speed and acceleration FINE tunning).
Perhaps setting the lookahead lines to 200 from the MACH3 default of 20 can also help.
Best Regards.
EDIT; Argumentative / contradictory comments removed by moderator.
-
Craig, Will you explain what is a ESS? Is that a breakout board?
-
Hi,
sorry, didn't mean to confuse.
ESS stands for Ethernet SmoothStepper, an external motion controller made by a US company Warp9TD.
There are a number of manufacturers including some highly regarded US and European Manufacturers and some widely criticized
Chinese manufacturers. Leaving aside the quality of the device and the support for it...what do you get with an external motion
controller?
Mach, the Windows application program we see is a Gcode interpreter and motion planner. The planned motion is a string
of numerical data describing the position of all axes in 1 ms time slices. A PC is good at crunching numbers as so it can decode
Gcode and plan motion very easily.
What happens to the numerical motion plan?
In the case of a parallel port machine the numerical data is transferred to Machs pulse engine, a block of code living in
kernel 0 level of the CPU. It calculates the required pulse streams and produces them with interrupt driven timers.
A PC's CPU is not optimized for this. Firstly it almost has to shove Windows out of the way and Windows doesn't like it
and fights back. Secondly Windows uses the CPU's interrupt system extensively for its own purposes. We cannot use the system
exclusively for our timers because Window would crash pronto. Lastly a PC's CPU does not have dedicated hardware timers
and so there is a software overhead associated with an interrupt driven timer.
The programming community all said 'it cant be done' and yet Art Fennerty did it and gave hobbyists a cheap (free) motion
controller and that's responsible for the huge popularity of Mach3.
About 10 years ago several manufacturers started making external controllers. They take the numerical data from Mach's motion
planner and generate the required pulse streams for the motors. They all use FPGA's, or microcontrollers or DSP chips or
sometimes all three. All of these devices have sophisticated built in hardware timers and can produce very fast and accurate
pulse streams well in excess of what Mach's pulse engine can do.
What do you get for it all? Firstly the PC's CPU is unloaded of the burden of producing pulse streams and therefore runs Mach
without the stuttering and stalling that often plagues parallel port machines. It means you can run 32 or 64 bit PC's (cf
32 bit only for parallel port), any Windows OS including Windows 10 (cf Windows7 and earlier only for parallel port) and you can
use laptops or desktops (cf desktop only for parallel port).
External motion controllers are faster than a parallel port, the ESS for instance can produce pulse streams up to 4 MHz,
compare that to Mach's parallel port of 25kHz (default). It means you can run high resolution servos and steppers that you
can't with a parallel port.
The ESS I mentioned is connected to the PC by an Ethernet cable and it in turn connects to a breakout board. at the current time
an ESS is about $190US. There are much cheaper units but all the good (US and European) ones are in the same price
bracket, ie not cheap. There are some rubbishy Chinese ones cheap....but they don't work well....if at all.
I used Mach's parallel port on a dual core Atom min-ITX board PC for a couple of years. It has one built in port and I added another
on a PCI card. It worked really well. You could make it stutter and stall if you were silly but by and large it was pretty good.
Then I bought an ESS. I could now run my steppers 33% faster without losing steps and they seemed to run cooler and
'sweeter' than before. I didn't expect that sort of improvement. I would have sworn that then parallel port was extracting
the best from my steppers but the SmoothStepper is just that...smooth. Mach is very much less inclined
to stutter or stall if some other program is running or it needs to compute a redraw.....and redraws often bugger up
the parallel port.
I apologize if it seems that I'm trying to 'sell' you another piece of equipment but my experience is that the simplicity
and performance that a good external motion controller offers is just light years ahead of a parallel port.
The chances are that an ESS, or other good quality external controller like a UC300 or 57CNC, are cheaper than
a new PC......and you wont have any bother trying to get a parallel port to work!
Craig
-
Thanks for the information. I've try every setting that has been recommended and still have problem. So i guess i'll be looking at other options until i can get the problem solve.
Thanks
-
can you post your xml
-
Can you explain what xml stands for?? I would be glad to post.
thanks
-
in c:\Mach3 is a file "yourprofilename".xml where all Settings are stored.
-
Here is my XML file.
-
had a quick look to your Profile. only x-axis is set.
Dir Port1 Pin3 LowActive1
Step Port1 Pin4 LowActive1
Step/per 3850
Velocity 103.98
Accel 4
Dir Pulsewidth 0
Step Pulsewidth 0
These two should be set to 5 see reply#30 from reuelt
-
had a quick look to your Profile. only x-axis is set.
Dir Port1 Pin3 LowActive1
Step Port1 Pin4 LowActive1
Step/per 3850
Velocity 103.98
Accel 4
Dir Pulsewidth 0
Step Pulsewidth 0
These two should be set to 5 see reply#30 from reuelt
Thanks TPS for great trouble shooting.
I have edited the XML file as the PO probably did not enter the <enter> key and save each axis.
@Ricky
Please try using attached edited XML file I have edited ( I have only enabled X, Y and Z) as starting point for further fine tunning. Please use 7zip to unzip.
Happy CNCing
Reuel
-
I will give it a try.
Thanks
-
I finally was able to down load the xml file and had to do a lot of fine tuning. The x axis runs very good on the mach 3 demo. The y axis was harder to tune and it also will run on the demo, but when trying to run both the x and y on the demo they start growling and sometimes stop. I have tried two different computers and a second breakout board. still have the same problem. I don't know what to try next. Any help is appreciated.
-
so what are your final Settings?
how did you do the X/Y test, via Input in MDI line?
-
Two items that I did not see mentioned that you might try.
One is to invert the pulse signal ... just for the heck of it. Some drives like it ... some don't.
The other is the voltage of the port. Possibly 3.3v ?
-
Here is the final settings: X settings
Y & A axis settings: steps 7100
steps: 6940 velocity 60
velocity: 59 accel 8
accel: 8 step pulse 5
step pulse: 5 dir pulse 5
dir pulse: 5
I tested by just running in the mach 3 demo and putting into jog mode. I will check the voltage at port.
-
If you'd care to try inverting the pulse signal, it is done here.
May, or may not help but usually does make a difference.
Russ
-
my voltage from ports
port 2 ( 3.2) 10 (3.4) when I hit the arrow key in opposite direction, only ports 3.5.7.10
3 (3,3) 11 (5) went to 0 if that means anything.
4 (3.3) 12 (5)
5 (3.3) 13 (5)
6 (3,3)
7 (3.3)
8 (3.3)
9 (3.3)
-
Tried the inverting, want move at all.
-
x running by itself sounds and does good.
Y and A by itself sounds ok and runs ok but not as good as the X.
Will barely run together and sometimes Y and A will stop and X will run very slow.
all three motors are long 34 steppers. Is the 3.3 volts from the computer port correct?
-
The docs for your drives specify a min. of 5v.
Russ
-
As Russ has mentioned, if your drivers require 5 Volt TTL signals and your Parallel Port is only outputting 3.3 Volt TTL then your CNC machine will never work correctly.
You could perhaps add a PCI Parallel Port card which outputs the required 5 Volt TTL – there is a guide here; https://www.machsupport.com/forum/index.php?topic=34848.0
Or
AsCraig has suggested, buy and use a suitable motion controller.
(there are BoB's and converters that will transform a 3.3V TTL to 5V TTL but they are more of a patch than a solution).
Tweakie.
-
will buy ESS and will post when connected to tell how it works.
Thanks
-
has been awhile posting but got the ESS installed, set motor tuning to x and y/a. can hold down arrow keys for x and y and motors run ok but when running demo they are doing the same thing; dragging, growling stopping. Did not help by buying ESS. Changed computers, no change. It doesn't do it at any certain line in gcode. I am very new at this and i'm sure I have missed a setting. Thanks in advance for any help.
-
Hi Ricky,
Could you perhaps connect a voltmeter to the output of your motor drivers power supply and monitor the voltage when the stalling / growling occurs - is there a voltage drop ?
Tweakie.
-
Hi,
I presume that you must be using a breakout board with the ESS? Is it the same breakout board that you used previously
when trying to get the parallel port to run?
The pulse streams to your motor drivers are generated within the FPGA of the ESS. I think you can rest assured that the
ESS is generating high quality pulse streams.
The problem therefore must be in those components that remain, namely the breakout board and the drivers. Do you have
or have access to an oscilloscope?
When you are running your machine you say it runs OK when jogging but fails with Gcode? Have you tried MDI instructions?
Craig
-
Here is the final settings: X settings
Y & A axis settings: steps 7100
steps: 6940 velocity 60
velocity: 59 accel 8
accel: 8 step pulse 5
step pulse: 5 dir pulse 5
dir pulse: 5
why did you not use the recommended 8128 steps/per ?
-
I will try and test the volts.
I did buy another BOB and not change. tried the MDI but the motors where very slow and rough sounding.
I did use the 8128 but could not get motors to run very well. I tried those numbers after the ESS install and had to go somewhere around 7000 steps, velocity 60, accel 8 . set the dip switches as recommended. still same problem. If I understand the ESS it does allow me to use 64 bit instead of a 32 computer and did upgrade to 64. should I go back?
I do not have a oscilloscope.
-
the 8128 steps/per is a value witch is given trough your microstepping and Resolution of your ballscrew.
in other word's it is fix and no parameter to Play with.
what exact motortype are you using?
what bob are you using?
-
I an using nema 34 and the breakout board db25-1205. Ok ill set that back to 8128
-
I an using nema 34
ok i was asking for a exact Motor Label/number, because nema 34 is like you say "i drive a green BMW",
there are hundreds of nema 34 brands, and i wanted to doublecheck the Motor wiring.-> waiting for reply
and the breakout board db25-1205.
here i found a manual
Ok ill set that back to 8128
just to clarify this will not solve the Problem, but the 8128 steps/per is a given value based on your Setup and Needs not
to be modyfied.
-
This is the label on motor:
Longs Stepper Motor
model: 34HS5435C-37B2
3.5A (U) 1.8deg/step
-
ok lets start again, i only want to double check everything.
Motor 34HS5435C-37B2:
red -> DM860A A
green -> DM860A A-
yellow -> DM860A B
blue -> DM860A B-
DM860A:
dip switches current
SW1 -> OFF
SW2 -> OFF
SW3 -> ON
dip switches micro stepping
SW5 -> ON
SW6 -> OFF
SW7 -> ON
SW8 -> ON
wiring signals
PLS+ -> DB25-1205 pin 2
DIR+ -> DB25-1205 pin 3
PLS- -> DB25-1205 pin GND
DIR- -> DB25-1205 pin GND
(only for the x-axis)
V+ -> Power Supply +
GND -> Power Supply GND
BOB DB25-1205 minimal wiring for x-axis
pin 2 -> DM860A PLS+
pin 3 -> DM860A DIR+
pin GND -> DM860A PLS-
pin GND -> DM860A DIR-
pin VDD -> external power supply 5V
pin GND -> external power supply GND
JN1 connected
JN2 connected
JN4 here i am not realy sure what the manual is trying to tell me
ethernet smoothstepper minimal wiring
25pole ribbon cable to db25 connector for BOB
ethernet cable to pc
external power supply for 5V connected
motor tuning settings (we discused a lot)
steps/per 8128
vel 60
acc 12
hope i got the mayor things, pls take the the time to go through this list step by step and doublecheck,
i have taken the time to create it.
-
I took the list and went one by one, I had the number 1 switch turned on so turned it off. the only thing I did not know about is the JN1 and 2 so unable to check that. All else match the list. Tried running only the x axis in the jog mode. Did not sound good. Than ran x in the demo mode, still stalling, more so when starting and changing directions. When running the motor sounded good other than the stalling. Did the DMI and setting was 8128 matched the list. I did try changing the vol and acc but did better at 60 and 12. Changed drivers no improvement, change BOB no improvement. Checked connection no problems that I could find. tried checking the power when in stall but not enough hands. will do that with help.
-
can you post a Video with Sound?
-
Checked connection no problems that I could find. tried checking the power when in stall but not enough hands. will do that with help.
To add to my earlier post...
You need to be looking for a voltage drop when the motors start growling or at the instant of motor stall.
Tweakie.
-
I will try to take a video.
Can you tell me exactily were to test for the voltage drop??
-
Can you tell me exactily were to test for the voltage drop??
Output of the power supply feeding the stepper motor drivers.
Tweakie.
-
power from power supply to driver and hope you can hear it.
-
Hi,
there is a small variation there but I would have thought well within bounds.
One possibility is that if when an axis, X say, is accelerating it will draw significant current from the supply and that will
cause a variation in the supply voltage. The variation could be construed by the Y axis driver as an input pulse and it would
therefore react. You could say that it is an example of 'conducted noise'.
Some people try to use the one high voltage supply to provide the stepper drivers but also feed a voltage reducer to provide
low voltage for your controller and BoB. It is not recommended as any variation in supply voltage is likely to be impressed on the
low voltage thereby introducing noise into your controller/BoB.
Do you have a separate low voltage supply for your controller/BoB combination?
Craig
-
Yes I do have a separate power supply for the BOB.
-
I do have power supply for each driver, and ESS and BOB.
-
is there a possible fix for this problem?
-
Hi,
it is far from certain what the problem is, there may in fact be two or three problems which are compounding things and
making it hard to diagnose.
You have bought yourself an ESS which is a good step to ensure good pulse streams to your BoB and stepper drivers,
yet the problem persists. It is not impossible that there is a problem with Mach/ESS/BoB but being able to eliminate
Mach/ESS/BoB altogether is the only real way to test the capability of the stepper/driver combination.
I suspect, but cannot be sure or prove, that the issue is with the steppers and drivers.
Electronics is my thing and I have a digital signal generator. Were this my installation I would drive the step input
of a driver direct from my signal generator. I could then be ABSOLUTELY certain that Mach/ESS/BoB have no part in the
test.
One of the challenges of testing steppers and their drivers is that it is not possible to just turn on the signal generator
and expect the stepper to be able to accelerate instantly. It is necessary to start the generator at a low frequency,
maybe a few Hz, and then increase the frequency taking note the frequency at which the steppers start to lose steps
or stall. The frequency at which the steppers stall is the uppermost limit that your steppers will perform. In a CNC machine
there will be loads induced by cutting forces and the torque necessary to accelerate the mass of the axes. Thus the ultimate
top speed that you have determined may be much less in practice.
Do you have access to a signal generator that could perform such a test?
The number of posts and the duration of this thread suggest to me that you must be getting very frustrated with the whole
thing. It might be time to enlist, even pay for, the help of a electronic technician. A stepper/driver test like this should not
require much more than half an hour for a technician. You really need to nail down where the fault is: either the stepper/driver
OR Mach/ESS/BoB.
Craig
-
hi,
one idea is left. can you pls do a
G1 X0.1968 F1
in MDI, and confirm that the Motor does exactly one Revolution?
-
I will have research for a signal generator and a electronic tech.
When entering the G1 xo.1968 f1, it keeps turning never stops and is very slow and rough sounding.
-
When entering the G1 xo.1968 f1, it keeps turning never stops and is very slow and rough sounding.
have you zeroed x-axis before Input?
-
no it was not. After zeroing it out, it did make one revolution.
-
ok that means microstepping and steps per are correct.
next step would be to have a look with oszilloscope for the step/dir signals.