[johnno402002]

Hi, I'm starting a new thread to discuss some issues I've been having with the spindle on my new SVM2. I want to say right from the outset that I'm still very impressed with the machine, and that I think that the problems I'm seeing are more than likely related only to my machine. So don't jump to any conclusions that what I'm seeing is a rerun of the Mikini saga, it's not.

So, there are two issues I've been trying to get to the bottom of,

One is that my spindle was having trouble rotating under no load below 500 rpm, and it had low torque at that speed, and it would stall completely at 300 or so rpm. These rpms are the S entry in mach3.

The second is that every now and then, the spindle speed would drop down in revs for a few seconds then recover.

Now, I'm a newby at setting the correct speeds and feeds, and I have a lot to learn, so I have been using gwizard to suggest the correct ones and after the first couple of incidents I've been using the feed rates rather conservatively. What I found was I could not get anywhere near the feed rates suggested, and I almost always had to increase the spindle rpm to get the chip load down. So I started using the endmill manufacturers calculations and I found they generally lined up with Gwizard, which left me wondering if the machine was actually doing what the G code told it to. The short answer, is no, it was not doing what it was told to do.

So, I spent a day reading the mach3 manual and checking what settings had been put into mach3, to make sure they were still what the manufacturer supplied, and they made sense. They were still the same, and they made sense, as best I can tell anyway.

There were several other issues that appeared once I started running some fairly basic tests. The most alarming was that when in mdi mode and running the spindle at or below 500 rpm random characters started appearing in the input field. These seemed to include the enter key, as they would be submitted as command and mach3 would give error messages. Then the home button was pressed ( not by me, from who knows where ) and the machine homed itself unexpectedly. The other issue is that the spindle developed a knocking noise, if it was a petrol engine I'd say that it had a big end knock. I took the belt off the motor, which confirmed that the noise was coming from the motor, not the spindle itself.

[johnno402002]

The good news is that quite a few of these issues have been resolved, but not all of them.

We decided to have a look at the signals coming back from the speed sensors on the motor. While connecting the dso one of the connections was found to be a bit loose. After connecting the dso, the motor and drive performed completely differently. The knock was gone, the random characters into the mdi field were gone, and we could get a lot lower spindle speed and it seems to have a lot more torque. I have yet to test it under load in a cut, but it is definitely running a lot better than it was. Here's a picture of what the signals looked like, plus the input to the drive from the breakout board, at about 500 rpm.

Attachment 315072

So that leaves a couple of issues, one is that the spindle is still not running at the speed commanded in Gcode, and the other is the drop in revs every now and again.

[johnno402002]

You can see the level of noise on the 0-10 volt input into the drive is quite high compared to the 1 volt base level at around 500 rpm. So, to see if that was having any effect on the drive and motor, we replaced the input signal with one from our lab power supply which has only 10mV or so of ripple. The result is quite amazing, we got down to about 105 rpm, and it still had lots of torque at 120 rpm. Video here. I should add that the sound on the recording includes the sound of the cabinet fan, the spindle actually sounds nicer than the video would suggest, or my ears have packed it in as well.



So, the conclusion we are drawing from this it that the drive is being impacted by the noise on the incoming signal, where it is coming from I don't have a clue yet.

[mcphill]

I would start by making sure you have good ground connections everywhere, and that all electrical connections are tight. Measure resistance between some grounds and grounds to the machine frame - they should ALL be near zero.

[johnno402002]

One thing we did notice, is that using the power supply we had very fine control over the voltage going into the drive. We noticed that the output of the drive goes up or down in steps, we didn't accurately capture this, but it seemed to be around 50 rpm steps at the rpm were we noticed it. So, small changes in the input can have no effect on the output, or, if you happened to be just below or above the the point where the drive steps to the next speed, they can have a significant impact on the output. As I understand it, the output from mach 3 is also in steps, and I'm guessing they are not the same steps as the input to the drive. So you could end up with a harmonic problem here where both the incoming signal to the drive and the internal drive hit steps together. I haven't seen any evidence yet that this is happening, hopefully it wont.

I think these step changes, in conjunction with the noise, is the cause of the change in rpm I see every now and again. Since the initial connection problem has been fixed, the magnitude of the change seems a lot less, and I think if we can get to the source of the problem with the noise these changes will just be a normal part of the drive behaviour where it changes steps.

Next step is to try to locate the source of the noise, or at least try to get an understanding of the circuit producing it. We have experimented with putting various capacitors across the output on the breakout board, but we had no success. Anything that quietened the noise moved the base signal level as well, and basically made it perform worse.

The machine is usable as it is I think, although I still have a fair error between the s value in mach 3 and the set value that gets put on the drive, This seems to be set on the spindle pulley config, the entries there currently are 0 for the lower limit, 6000 for the upper limit, and 1 for the ratio. I have experimented with changing these and found that if I put 300 in the min setting the drive will fault. If I leave it at zero, and drop the max to about 5200 rpm the correlation between the drive setting and the s setting seem closer at some rpm, but nowhere near correct.

I will have to try cutting under load to see if it performs any better than it was.

John

[johnno402002]

I would start by making sure you have good ground connections everywhere, and that all electrical connections are tight. Measure resistance between some grounds and grounds to the machine frame - they should ALL be near zero.

Will do, thanks mcphill,

Your posting of Defeng's instructions in the magnetic drawbar thread was very helpfull, can we put it in this thread as well, it might be bit easier to find here?

[johnno402002]

I've found that the ground wire from the motion board under the breakout board, to the breakout board is not earthed. I'm not absolutely sure that it should be, but all three wires that are connected to the drive to control the speed are floating.which seems to explain the noise. I have the manual for a version of the JNC40M motion board, it says it is a simulated circuit which should be connected to the ground on the frequency changer. I'm translating that to mean its not a real ground but it should be connected to the motor driver ground pin associated with the speed voltage, which it is.

And just for fun, the hard drive in the PC running mach3 just died :tired:, so now to rebuild the pc before I can go any further.

[ger21]

We noticed that the output of the drive goes up or down in steps, we didn't accurately capture this, but it seemed to be around 50 rpm steps at the rpm were we noticed it

The Mach3 manual explains how to determine the size of the steps, if your using PWM.

[johnno402002]

The Mach3 manual explains how to determine the size of the steps, if your using PWM.

Hi Gerry, thanks for that. I've been trying to get that bit working correctly as well, but that was not what I was referring to here. What I meant was that the drive itself is operating in steps, it did not react to the changes we were making on it's input voltage using the lab power supply, until the change was big enough to cause it to jump to the next step. Mach3, at this point of the test was disconnected from the drive..

But, on the subject of mach3 steps, I was reading a manual for the motion card yesterday and it has a screen shot of mach3 where you select the mach3 frequency. In the screen shot it shows 25khz, in the text it says 100khz. Mine is set at 25khz. The 25khz obviously makes the steps bigger, I'm thinking I should change it to 100khz and then make sure the axis drives are still tuned properly. What frequency do you run mach3 at?

John

[ihavenofish]

25khz is the baseline and most reliable setting for mach3. 100khz is the maximum, and may or may not work well on your PC. you will need to test it thoroughly. some things on some break out boards such as charge pumps "expect" 25khz and wont work with any other setting.

as for the steps...
the drive is digital. it is converting the analogue input to digital. if its an 8 bit ADC then you get 256 steps in speed. if its 10 bit you get 1024 steps etc.

on the mach 3 side, the pwm duty cycle is the same. 8 or 10 or 12 or 16 bits (not sure which).


in the end none of this matters much as you are setting a speed and leaving it there. you dont generally need anything more precise than 50rpm steps on a spindle. you arent usually going to cut under 500rpm in practice id expect.

what is important is that you have full torque and fast response to load at all speeds. with the changes you made to the loose wire you said it improved vastly. id like to see it taking a nice heavy cut now (at any speed). this will give us a big hint on fixing other drive/motor sets.

[hab82]

Hi John
I've tried to track down which model of the SV2 that you have. In looking at your thread it doesn't sound like the one I just ordered. I know that I went with the AC Servo spindle to avoid some of the problems that the Mikini 1610L had. I am hoping that I will not have the same problems.

[ihavenofish]

Hi John
I've tried to track down which model of the SV2 that you have. In looking at your thread it doesn't sound like the one I just ordered. I know that I went with the AC Servo spindle to avoid some of the problems that the Mikini 1610L had. I am hoping that I will not have the same problems.

ac servo would not have this problem. its only related to bldc "spindle" motors.

[hab82]

Thanks.
That is why I chose the AC servo. I wan't to use it for tapping and like lower speed.

[mcphill]

Will do, thanks mcphill,

Your posting of Defeng's instructions in the magnetic drawbar thread was very helpfull, can we put it in this thread as well, it might be bit easier to find here?

Sure Thing!

Hi M,,

I just made a briefing of the BLDC driver. It's simple and should easy to understand.
Attachment 276162
BLDC DRIVER program set:
1. press "set/enter" ->F0 press "set/enter again"->show1401->press "UP""DOWN" to show "1111"->press "set/enter" to unlock the program.
2. press "UP" "DOWN" to select a function to set parameters.
3. "F4" to set MAX speed. your driver should set 5000 or 6000. after adjust the number, press "set/enter"to confirm your input.
4. "F6" to select the speed control signal source.
0=speed control by "UP/DOWN" keys on the panel;
1=speed control via 0-10V or 0-5V analoge voltage. Don't use both together!
2=speed control via PWM duty. 0-95% refer 0-max speed.

5. "F8" low speed range P parameter. "10-100" may be the best range. if P is too big, may cause over current when accelarating. if P is too small, motor will accelarate slow and torque is weak.
6. "F9" low speed range I parameter. 10-150" may be the best range. turn I bigger, the motor will be more sensitive to speed changing. turn I smaller, the motor will response slow and may cause speed not stable.
7. "F10" high speed range P parameter.
8. "F11" high speed range I parameter.
9. "F12" exit program set.

Don't have to re-power up the driver to make the parameters effect.

I hope I listed everything clearly and right.lol Any detail questions can ask me directly via email also.

Defeng
Here is Defeng's instructions on MY motor controller. Not sure he is using the same one on his machine or not, but if you have a BLDC spindle motor, my guess is this is at least VERY close to the driver he used...

[johnno402002]

Thanks mcphil, very handy info.

hab82, I see your questions have already been answered. I'd just like to add that I think the problems I have with mine may not be a good indication of what these BLDC motors and drive can do. When I ran the input to the drive from my power supply I got enough torque down around the 120 rpm mark to do tapping I think. I have the computer rebuilt, and I hope to track down the source of the noise this weekend. It could be coming from the pc I'm using to run mach3, down the usb cable, it may not be anything at all wrong with the machine.. I will do some test cuts once I reduce the noise. I cant help thinking there must be a way to measure the torque, but everything I think of is too complex and expensive, or too destructive. I will be very interested to hear how your ac drive performs, it wasn't an option when I ordered mine.

[johnno402002]

Hi,
I've done a couple of tests on the drive circuit to try and figure out where the problem might be. I've calculated the required voltages that need to be applied across a range of spindle rpm, based on zero volts is zero rpm, and 10 volts is 6000 rpm. I've checked that the mach3 pulley ratio settings are correct.


The circuit goes something like this Mach3 -> usb driver -> usb -> jamen motion board -> breakout board -> spindle drive -> spindle

The first test, I just set the rpm using the S command in Mach 3, and read the set rpm from the drive display across a range of rpm settings

The second test, I disconnected the connection from the breakout board to the drive, and hooked my power supply directly into the spindle driver. I used the same range of rpm settings, applying the calculated voltage to the drive. I've graphed the results.

[johnno402002]

So, what I'm seeing here is that above 1000 rpm, the drive is setting a higher rpm than the voltage applied. From mach3 the applied voltage to the drive is consistently lower the it should be. Below 1000 rpm, both the drive and the signal from mach 3 are lower than expected, Below 500 rpm it all goes to hell, which is why the motor has no torque. More tests to come, I'm going to break the circuit between the jamen motion board and the breakout board and inject my voltage there, to see if the breakout board is playing with the voltage. I suspect it is, I can see speed adjust written on the board, but no adjustment screw is visible. It may be underneath the board.

[johnno402002]

The breakout board does not alter the rpm voltage, it passes straight through. Where the rpm adjustment pot would have gone it is bridged out on the board.

[victorofga]

I experienced similar fluctuations with vfd... im not sure if related to the original poster issue..
changing the vfd fluctuation gone..
im not challenged so much that I could track down the issue..

[ihavenofish]

this isnt a vfd, and he cant really change it out as there is no other easy option.