Does this sound normal?

[grantx]

I have recently purchased a CNC with the following configuration:
2.2KW spindle.
DM860A controllers
KK01 interface board (from cnc4you.co.uk)
34Nema motors.

Jogging the machine across the x or y axis sounds sweet. No problems. When I run a programme though it does certain moves and the whole thing sounds terrible, I made a video:
https://youtu.be/aSbQEHxy5Uo

Does that sound normal to you?

I must also say that if I speed it up it almost eliminates that vibration noise but then it runs at a feed rate that would break bits.

[m_c]

Couple questions to start.

Does it still do it with the spindle not running?

Is the machine maintaining the correct position?
I.e. if you start at X0Y0, run the program, then command a move back to X0Y0, does it return to the same position.

[routercnc]

No, it does not sound right. I'm leaning towards electrical problem rather than mechanical, especially running at a higher feed rate removes it. Check all the electrical connections are tight.

Interesting that jogging is OK but running code is not. Are these both at the same feed rate? If you use feed rate override then jog the machine slower manually is it still OK?

Which software generated the g-code? Maybe paste it here in case there is something odd there.

[JAZZCNC]

Does that sound normal to you?

Oh, here we go again, I'm truly sorry to have to say this but for that particular machine then yes it's about right.!! That doesn't mean it's correct, it's very wrong.

The problem and the noises you are hearing won't be from one thing, it's a mixture of things. I have had plenty of experience with these machines bought from eBay and they are so badly engineered and constructed that a good percentage of the noise will be from poor alignment and setup of the ball-screws.

The next issue is that they are never set up regards motor tuning and micro-steps etc, don't know if you bought it directly from the guy in Peterlee or not but nearly everyone I've encountered hasn't come with Mach3 setup at all.

The third issue is that they use Nema 34 motors and often with a low voltage PSU which when combined with cheap drives or poorly setup drives they perform like a can of marbles due to resonance, this resonance is mostly what your hearing as the drives are either setup wrong regards micros steps or just not dealing with the resonance, which is also compounded by the poorly aligned and setup ball-screws. Resonance can and does cripple the performance of a machine and not to be underestimated.

The first port of call is to check what micro-steps are set on the drives.?
Do you know the voltage of the PSU feeding the drives.?

Is this using the UC100 controller.? ( This connects to the KK01 and goes to the PC)


I must also say that if I speed it up it almost eliminates that vibration noise but then it runs at a feed rate that would break bits.

What speed does it go away at.? Also, knowing these machines the reason they would break bits above 3mm diameter if cutting in wood's or MDF, etc is that they can not travel fast enough and the machine design is weak and poorly built which causes vibrations so together they can't reach the feed-rates to cut correctly.!

New users often believe it's their fault they are breaking cutters due to wrong feeds and speeds, often wrongly slowing down when they should be cutting faster. However, this machine in particular while being built from steel is so badly constructed, set up, and poorly specified then they are fighting a losing battle from the get-go.!

The reason for them not being able to reach the feeds is two-fold, 1st they only have 5mm pitch ball screws, and 2nd they have Nema 34 motors with low volts.
Large motors like Nema34 when run on low voltage don't spin very fast and the torque drops away very quickly when they are at higher rpm's, they struggle to reach much above 1000rpm and have any usable torque left if run on voltages lower than 60Vdc.
Combine this with a 5mm pitch and your maximum velocity is 5000 mm/min, which in practice would actually be more like 4500mm/min to avoid losing steps.
However, This would be your Rapid feed rate and wouldn't be of any use for cutting as the torque will be quite low. So your usable cutting velocity would be more in the range of 3500mm/min which simply isn't fast enough to cut most woods or MDF correctly with any tools above 3mm.
Combine this with the vibrations and flex from the machine and you are always going to break cutters and struggle to get a decent finish on the material, also tool wear will be excessive due to wrong cutting parameters causing excess heat.!

Again I'm very sorry to have to point these things out and these machines make my blood boil because I see people getting ripped off all the time with these poorly built, poorly specified, and set up machines with no support from the supplier and it's not fair.
However, I don't know how I can let people know what they have and are facing without being blunt and to the point informing them of what they have in front of them. We can help you regards setup etc and possibly getting it running a little better but the rest isn't anything we can help with other than to advise you to sell it rather than try to upgrade or improve it.

[m_c]

After Jazz's post, I've just realised who's machine it is.

I was thinking it was somebody's secondhand home built machine you had purchased.

[JAZZCNC]

Watching the video again I can hear as the machine stops a thud noise, can also hear it in the movement.! This sounds very much like one of the ball-screws is floating in the BK end bearing which causes backlash. So check that the nut hasn't come loose which can be common on the cheap BK bearings which are fitted on this machine.

To check for end float grab the screw firmly and push/pull towards the bearings, there should be NO axial movement.

But you also have issues with micro-step setup and motor tuning as well which is most of the strange noise (resonance)

[JAZZCNC]

After Jazz's post, I've just realised who's machine it is.

I was thinking it was somebody's secondhand home built machine you had purchased.

MC You can easily spot them because every other bolt is missing from the rails..:hysterical:

[Ollie78]

Doesn`t sound good.

Did you buy it new ?
If so, surely you should be able to send it back or at least get it repaired under warranty or something.

Ollie

[dazp1976]

Oh, here we go again, I

The reason for them not being able to reach the feeds is two-fold, 1st they only have 5mm pitch ball screws, and 2nd they have Nema 34 motors with low volts.
Large motors like Nema34 when run on low voltage don't spin very fast and the torque drops away very quickly when they are at higher rpm's, they struggle to reach much above 1000rpm and have any usable torque left if run on voltages lower than 60Vdc.
Combine this with a 5mm pitch and your maximum velocity is 5000 mm/min, which in practice would actually be more like 4500mm/min

Even 60Vdc isn't enough to reliably run 34's in my view.
I know I'm talking of a mill with gibs here and routers are looser running, but still., I could not get my 1080oz reliably above 450rpm (2250mm/min).
I've bought a new driver for that particular motor (110vdc/80Vac) and upping to 80Vdc as a tester to see if there's any increase in performance. (Will be a while yet, still building new control box).
Will eventually go all AC.

With the X & Y I swapped motors to 566oz nema 24's keeping voltage 60vdc and 860 drivers the same.
It's twice as fast with those 2 on there.

It's shocking how many people get screwed over these days with misleading specs and advertising.
At least if you D.I.Y. it you can over engineer the damn things!. (you can also go too far there though).

[JAZZCNC]

Even 60Vdc isn't enough to reliably run 34's in my view.
I know I'm talking of a mill with gibs here and routers are looser running, but still., I could not get my 1080oz reliably above 450rpm (2250mm/min).
I've bought a new driver for that particular motor (110vdc/80Vac) and upping to 80Vdc as a tester to see if there's any increase in performance. (Will be a while yet, still building new control box).
Will eventually go all AC.

Yes, I agree but in this case, with small Nm NEMA 34's like fitted on this machine then 60Vdc or more should get them to close 1000rpm. It's a combo of 5mm and low rpm which cripples the speed.


Will eventually go all AC.

Don't get hung up on AC, the only advantage is you don't have to mess around building DC PSU's. Even with AC power the drives convert this to DC for the motors.

[grantx]

Well, that is a very discouraging read.

Yes, I bought it new from David Ward in Peterlee. It is less than 2 weeks old and he's not returning my calls anymore (he probably found this thread).
I've tried attaching a few pics of the machine but I cant attach more than one for some reason.

His advert on Ebay says I can return within 2 weeks but the problem is that I didn't pay him through ebay - it was done privately.

The power supply is a Meishiie S-1200 -60
Output 60V and 20A.
Drive controllers are all DM860A

Frame seems very sturdy. There is no play or axial movement between rails and ball screw blocks.

What do you think is the best way forward from here?


30099

[grantx]

Yes the machine maintains the correct position. If I ask it to move 25mm it moves exactly that. Returning to home is accurate as well.

[grantx]

.

Yes, it uses the UC100 controller.
Jogging the machine across x and Y and there is no sound. Ive attached my settings from Mach3 that I am using. But these are different to what is run in the Gcode. In that, I am using 1800mm/min.

X axis:
Steps per: 125
Velocity mm/min: 3600
Acceleration: 500
G's: 0.050
Step pulse: 3
Dir pulse: 3

y axis:
Steps per: 125
Velocity mm/min: 3700
Acceleration: 500
G's: 0.050
Step pulse: 3
Dir pulse: 3

Z axis:
Steps per: 125
Velocity mm/min: 1000
Acceleration: 500
G's: 0.101
Step pulse: 3
Dir pulse: 3


A axis:
Steps per: 125
Velocity mm/min: 3600
Acceleration: 500
G's: 0.050
Step pulse: 3
Dir pulse: 3

[JAZZCNC]

Well, that is a very discouraging read.

Yes, I bought it new from David Ward in Peterlee. It is less than 2 weeks old and he's not returning my calls anymore (he probably found this thread).

Like I said I'm very sorry to have to say it and it's wrong that should need to. Also, I don't think he's not answering because of reading this thread, I've had others contact me privately and he doesn't answer their calls either.!


His advert on Ebay says I can return within 2 weeks but the problem is that I didn't pay him through ebay - it was done privately.

If you got a receipt then shouldn't matter how you paid him. I also believe you have 30days under your statutory rights to return the machine, but I'm no expert so check it out.



What do you think is the best way forward from here?

If it was me then I'd be returning it without a doubt but only you can make that decision.

[the great waldo]

Anything bought online can be returned in 2 weeks. So be quick with a return claim. Cheers Andrew

[JAZZCNC]

Ok well, straight away I can see something is odd because 125 Steps per setting doesn't add up.? The DM860A doesn't offer a micro-step setting that matches this setting when used with a 5mm pitch ball screw driven direct from the motor like what you have. 5 * 125=625 and the nearest micro step setting in the manual I'm looking at is 640.

This is even more confusing if you are 100% sure that it moves the exact amount because it shouldn't unless the drives are using a different setting to what the manual I'm reading says or the ball screw pitch isn't 5mm.?
Both these are easily checked, the drives will show the dip switch settings on a sticker and you should be able to match the dip switch positions on the drives. To find the screw pitch then just jog or manually turn the screw one revolution and measure how far it moved.

However, If I'm correct and the pitch is 5mm and you are correct that it's accurately moving the distance told then at best the micro-steps setting must be 625 and this could be the problem.?
These drives I think are the cheap nasty analog version and they will at best only have basic resonance handling built into them, Analog drives are notoriously bad at handling resonance and to be honest, should be gone the way of the dodo by now.!

You will find if you change the micro stepping on the drives to 1600 that the motors will run a lot smoother, you will also need to change the Steps per in the motor tuning to match.
This is done by dividing the Micro steps by the ball screw pitch ie: 1600/5=320.

Doing this will change how the motor operates and hopefully will shift the resonance into a range the drives can deal with. You will instantly hear the difference.
Analog drives are old technology and it was common to have to play with micro-steps to try and find a sweet spot, esp if used with cheap motors which these probably will be.?


All this said though your problem could still be a mixture of mechanical and low-quality electrical that is poorly set up so if you are thinking to send it back then I wouldn't touch it.
Also, if it does run smoother you still have a machine that isn't fast enough to cut at the correct feeds n speeds for most materials and while you think the frame is sturdy I can tell you it isn't really and the way it's constructed means it cannot be accurate which you will come to realise in time as you try to do more intricate work or complex work.

So again my strong advice is to try to return it because you'll will still have a lemon if you get it running smoother.

[grantx]

Here are a few more pictures.
5mm pitch lead screw
34 Nema
Drivers showing jumper positions (Off - On - On -Off)

301013010230103301013010230103

[JAZZCNC]

Here are a few more pictures.
5mm pitch lead screw

Measuring like that can be misleading because the screw could be a twin start so that would be 2 x 5mm starts making it a 10mm pitch. This is why I said rotate and measure the distance it moved as that is the more accurate method.

[JAZZCNC]

Drivers showing jumper positions (Off - On - On -Off)

Ok with those settings using the manual I'm looking at the micro-steps are set to 2000. This doesn't ring true to the mach3 steps per setting because it should be 400 if 5mm pitch and 200 if 10mm pitch. You have mach3 set at 125 Steps per which means you can not possibly be moving the correct distance unless the screw pitch is different to 5 or 10mm which I very much doubt.
The motors are 1.8degree so 200 steps per 360degree's and clearly direct drive so there are no ratios applied so that only leaves the screw pitch which can affect the movement and I'd be very very surprised if it's anything but 5mm or 10mm. This means it can't possibly be moving the correct distance unless the dip switch settings are wrong.

[grantx]

You are right. I dont know what I measured before....
1 revolution gives 15mm travel on the Y axis (I havent measure the others)

When I tell it to move 100mm (G0 y100) it moves 150mm.

EDIT: Ok I remember how I measured it before. Using the auto-calibrate in Mach3 I get it to move 100 on the y-axis and it moves exactly that. It tells me the steps are 125, So that must be correct.

I dont know what that G0 y100 command is doing ...

[JAZZCNC]

You are right. I dont know what I measured before....
1 revolution gives 15mm travel on the Y axis (I havent measure the others)

When I tell it to move 100mm (G0 y100) it moves 150mm.

EDIT: Ok I remember how I measured it before. Using the auto-calibrate in Mach3 I get it to move 100 on the y-axis and it moves exactly that. It tells me the steps are 125, So that must be correct.

I dont know what that G0 y100 command is doing ...

Ok well, 15mm isn't a common pitch size but they do exist. However still, things don't add up.?

The drives are set for 2000 micro-steps and your direct drive to the motors. 2000/15=133.333333 Steps Per mm.

If we work it backward just in case the drives dip switches are telling porky's then from the Mach3 Auto calibrate setting 125 steps per with 15mm pitch this would give a micro-step setting of 1875.? I've never seen drives that allow this setting and the manual doesn't show it either.?

So something isn't adding up here at all.

Regards the G0 y100 command then it will move to y100 coordinate position which doesn't mean it will necessarily move 100mm. If you Zero the DRO then it should move 100mm.
To move a set amount without worrying about the coordinate location then you need to switch to incremental mode, try this next time.
G91 G0 y100 and it will move the amount asked for from the current location. When finished type G90 to put back into absolute positioning mode.

[andy_con]

I have recently purchased a CNC with the following configuration:
2.2KW spindle.
DM860A controllers
KK01 interface board (from cnc4you.co.uk)
34Nema motors.

Jogging the machine across the x or y axis sounds sweet. No problems. When I run a programme though it does certain moves and the whole thing sounds terrible, I made a video:
https://youtu.be/aSbQEHxy5Uo

Does that sound normal to you?

I must also say that if I speed it up it almost eliminates that vibration noise but then it runs at a feed rate that would break bits.

Sorry to hear you have joined the club, its not a good club to be in.

http://www.mycncuk.com/threads/14563-Wood-Works-Ebay-cnc-router-review

[grantx]

Andy, did you get this resolved? I got in touch with him yesterday and he said he will be sending out replacement upgraded controllers which have much smoother output. We shall see. I will update this thread if I receive the kit.

I have been looking at the AXBB-E controller with closed-loop servos. Obviously, in hindsight this is the right gear for the job but it is my fault for not putting enough research into this. I just have to make the most of the situation and get it sorted as best I can.

[JAZZCNC]

Andy, did you get this resolved? I got in touch with him yesterday and he said he will be sending out replacement upgraded controllers which have much smoother output. We shall see. I will update this thread if I receive the kit.

Hi Grant, I know we have spoken and emailed so you know most of this but I'm saying what's below for the sake of others reading this.

It's not just as simple as swapping out the drives, motors and PSU also need to be factored into the equation.
I would want to know the specs of the motors and the inductance ratings, etc because usually, it goes hand in hand that if used cheap drives they will have used cheap motors probably bought together as a kit.
It's very common for Cheap NEMA 34 motors to have very high inductance which means more voltage is required to get any decent performance so just changing the drives won't always make much difference to performance.

Having a good machine is all about balance and how the components match up with each other and the machine in general regards ball screws etc.

BULLSHIT can baffle the unknowing or unweary so be very careful because you are about to walk into a minefield of BULLSHIT :cower:

[dazp1976]

Andy, did you get this resolved? I got in touch with him yesterday and he said he will be sending out replacement upgraded controllers which have much smoother output. We shall see. I will update this thread if I receive the kit.

I have been looking at the AXBB-E controller with closed-loop servos. Obviously, in hindsight this is the right gear for the job but it is my fault for not putting enough research into this. I just have to make the most of the situation and get it sorted as best I can.


Do you need to use more than 4 axis?.
I know of another guy who has the AXBB-E. It is a good unit however.....
If you want to use a 5th or 6th axis you need another breakout board to go into channel 3's non isolated IDC port.
The optocouplers on the isolated channel 2 outputs aren't fast enough for pulsing by the looks of it.
He got a UCSB board for chnl 3 and all is well.
Other than that they're decent units.


I had the 'kit' type electronics on my mill and almost the whole lot has been replaced due to lack of performance.
Went with the UC300eth, UCBB, and was going to switch to closed loop gear later on.
Closed loop is a little misleading though so I'm hoping to eventually change to Lichuan servo's instead.

[dazp1976]

BULLSHIT can baffle the unknowing or unweary so be very careful because you are about to walk into a minefield of BULLSHIT :cower:

Don't forget the minefield of crap electronics out there too!.
Even with stepperonline products you have to watch, they have good and bad. You really have to scutinise ALL the spec sheets.

[JAZZCNC]

Do you need to use more than 4 axis?.
I know of another guy who has the AXBB-E. It is a good unit however.....
If you want to use a 5th or 6th axis you need another breakout board to go into channel 3's non isolated IDC port.
The optocouplers on the isolated channel 2 outputs aren't fast enough for pulsing by the looks of it.
He got a UCSB board for chnl 3 and all is well.
Other than that they're decent units.

The motion controller is exactly the same one as the UC300, the only difference is the I/O board of the AXBB-E only provides 4 Axis outputs with provision for 2 more using an external breakout board. Other than that the motion controller side and performance are exactly the same. The only reason really to use a UC300 over the AXBB-E is if you need more I/O for things like ATC, etc.




Closed loop is a little misleading though so I'm hoping to eventually change to Lichuan servo's instead.

There is nothing misleading to the closed-loop stepper system, it does everything it says on the tin.? It's people misunderstanding the difference between steppers and servos which causes confusion.
The loop is closed back to drive using the encoder for error checking just like a Servo works and if the error threshold is exceeded the drive faults.
The BIG difference is in the power and performance of AC or DC Servo motors compared to Steppers, how they actually work regards closed loop is very very similar.

[grantx]

I only need the three axis. xyz.

Why do you say closed loop is misleading?

Hypothetically if I did need to upgrade, what would be a suitable setup? Now I do realise that the quality of cnc equipment lies on a very broad spectrum of quality standards and that there are unlimited combinations available, making my question nonsensical to seasoned veterans here. However.....in my situation where I will be cutting exclusively wood (soft and hard), possibly a few hours a day, on a sturdy metal frame with 15mm ball screws attached - what would a proper setup look like?

4 x Nema23 closed loop. (2 for X, 1 for Y, 1 for Z)
AXBB-E controller with accompanying software.
2 x Power supply ( 1 for X and 1 for Y and Z)
Controllers (Do I need 1 for each motor drive?)

[dazp1976]

The motion controller is exactly the same one as the UC300, the only difference is the I/O board of the AXBB-E only provides 4 Axis outputs with provision for 2 more using an external breakout board. Other than that the motion controller side and performance are exactly the same. The only reason really to use a UC300 over the AXBB-E is if you need more I/O for things like ATC, etc.


That's the plan (If I can get the machine to fund it itself). I'd love to have a go at making an ATC just for the fun of it. :smile:
Also. I prefer to have seperate components. Yes it can be a higher outlay but it's then cheaper to replace individual parts later on if one fails.

With closed loop..... I suppose it depends on what components you get. I've seen plenty of good/bad experiences over on cnczone. Mostly the HBS drivers.

[dazp1976]

I only need the three axis. xyz.

Why do you say closed loop is misleading?

Hypothetically if I did need to upgrade, what would be a suitable setup? Now I do realise that the quality of cnc equipment lies on a very broad spectrum of quality standards and that there are unlimited combinations available, making my question nonsensical to seasoned veterans here. However.....in my situation where I will be cutting exclusively wood (soft and hard), possibly a few hours a day, on a sturdy metal frame with 15mm ball screws attached - what would a proper setup look like?

4 x Nema23 closed loop. (2 for X, 1 for Y, 1 for Z)
AXBB-E controller with accompanying software.
2 x Power supply ( 1 for X and 1 for Y and Z)
Controllers (Do I need 1 for each motor drive?)

Ask JAZZCNC (Dean) about the Lichuan A86 drivers and 2102 motors he has if you're serious about closed loop.
One of the best options in the u.k. atm afaic.
You can hook those up directly to a 55vac toroidal power transformer for simplicity.
The AXBB-E has enough outputs for 4 axis, spindle and ancilliaries straight out the box (3+ slave in your case).
Initial outlay is a cost but..... you won't be constantly pratting about with a setup like this if you wire it correctly first time.

The documentation on the Lichuan is ok and I'm sure Dean will help you out.

[grantx]

Update on the resonance issues. David sent me replacement driver modules which I was initially very skeptical about however they have totally eliminated the problem. Ive run a profile which, prior to the change, was causing terrible vibration and noise but now runs totally smoothly. I can barely even hear the motors now.

I must say a huge thanks to Dean who was very supportive. He chatted with me over the phone and gave some excellent advice. I think Deans machines look of a very high spec and he obviously knows what he's talking about and as much as I would love to own one of his builds they are unfortunately a little beyond my budget at the moment.

A few have told me to send this cnc back but so far David has been very supportive and since my original complaint has been suitably resolved I don't have a valid reason to return it - nor would I want to. He has done an honest days job and I have what I paid for.

[JAZZCNC]

Pleased you are a little further forward and hope it goes well for you.:thumsup: