[sn6111]

Hello,

I have bought the 3020Z-D52, the new version with the black box, ballscrews and the couplings are spiral with set screw.
The machine is losing some steps and I don't know why happened this.
I replace the cables with shielded, I tried with different speeds but the problem hasn't solved.

I made an engrave, cutting in wood and the final result isn't good.
The letters are not straight, has inclination of a few degrees to the bottom.
The letters on the CAD are straight.
I have the same problem and with PCB engraving, cutting, drilling.

I'm thinking to buy the Gecko G540 but I don't know if the problem is from stepper driver.

Is any solution to do?

Thank you in advance
Nikos

[LeslieTroyer]

make sure all the stepper couplings are tight, cut back your accelleration, and maybe your velosity (do a test with 1/2 the accel and 2/3 the velosity to see if that helps ). what I did was "drill a hole" - route a large patch in both x&y then come back and drill the same hole. This will tell you if you are loosing steps under load. If you are having problems with the Z direction - do several minutes of routing with varying Z depth - and then recheck you zero is still where you expect it.

I have the same controller, and slowing down the accell really really helped.

Les

[sn6111]

Thank you Leslie,

The couplings are tight.
I will try and I will see the difference.
Could you send me the numbers of accel and velocity that you use on Mach3?

Your 3020Z-D52 CNC works very well after this change? No losing steps?

[LeslieTroyer]

I'll check tonight when I'm out in the shop.
Les

[sn6111]

Thank you Leslie.

I tried different velocity times, acceleration times but the problem still alive so I made an order for the Gecko G540 and I'm waiting to receive it.

Now I have one question, Could I connect the PWM controller from D52 (black box) to VFD output of the Gecko to control the spindle via Mach3 or I need a new spindle controller? The spindle that I have is 230 Watt DC Motor.

Thanks

[Rakounet]

Hi Nikos,

Some days ago, I purchased a similar model (I say "similar" because not YOOCNC branded). It is equiped with the same controller YOOCNC NT65-3X
My machine is also losing steps on the X-axis, but only when returning.
I made some tests to identify what was wrong. See the following video : Losing steps on CNC 3020Z - YouTube
Finally, I suspect the controller card which excludes the mechanical parts (rather a good news...).
I am in contact with the supplier. Hope to get a solution.
I will share it if it solves the issue.
Rak

[sn6111]

Hello Rak,

Thank you for your response.
I had the same problem with the X when returning.
I tried several methods but without any good solution.
Then I bought the Gecko G540 and I solved the problem.
No missing steps and very smooth running.

I think you will loose your time with the 3020Z controller. I understand , the Gecko is little expensive but is the only solution if you want a very good result.
The chassis and the motor are good but the stepper controller is for the garbage.

Regards,
Nikos

[Rakounet]

Hello Nikos,
Before buying a new controller, I have made some analysis with an oscilloscope and I have found the X signals were very noisy (more than Y and Z).
I have replaced C36 (electrolytic capacitor) by a new one and now, it works fine! I recommend also to add a 470 nF ceramic capacitor in parallel with C35, C36 and C37.
Hope it can help the community.
Regards,
Rak

[helisat]

Hi Rakounet

I have the same controller on my YOOCNC 3040. I too am missing steps at low speed, but at high speed it runs smoothly. Was this the case for you too?

I'll try replacing the electrolytic capacitors, and add the ceramic caps in parallel too. Hopefully this will fix it.

Cheers,
Luke

[helisat]

Update...

I replaced C35, C36 and C37. I also increased these to 2200uF (from 1000uF). Unfortunately it's still skipping occasional steps on the X-axis. The x-axis doesn't 'sound' as smooth as Y and Z.

To try some troubleshooting, I connected the Y-motor into the X-port of the controller, and the X-motor into the Y-port of the controller. Now the X-axis is okay, and the Y-axis was skipping steps. Now I know that it's definately the X-axis in the controller box that's the issue. I've analysed the controller PCB, and X, Y, Z circuitry all look identical. I then thought the 24v fan (next to the X-axis IC) might be causing interferance, so disconnnected it -- unfortunately still no fix.

Any other ideas?

[Alwaysintheknow]

If you haven't changed out the cables with the shielded type you are fighting a loosing battle.

[Rakounet]

Hi Luke,

Don't forget to add 470 nF ceramic capacitors between the power supply line and ground (pin 15 of each TB6560). The power supply line goes under the 0.5 ohm resistances (scape the vernish) as shown on the attached picture.

I don't have noticed any difference between low and high speed.

Let me know your results after adding the ceramic capacitors.

Cheers,

Rak

[helisat]

Hi Rak,

The local shop doesn't have ceramics caps as high as 470nF. Will a 470nF MKT polyester be okay?

Cheers,
Luke

[helisat]

Hi Rak,

Great news. I added the 470nF capacitors (MKT) and its all fixed. Everything is running very smoothly, and no missed steps.

Thanks,
Luke

[Rakounet]

Hi Luke,

Indeed, you can use MKT capacitors. I did it and it worked fine for me. The TB6530 datasheet recommends ceramic ones, that's why I indicated the ceramic type.

I have analysed the signals with an oscilloscope and I found the signals were very noisy, that's why I added 470 nF capacitors.

Have a look on the TB6530 datasheet and you could see the electrical design of the YOOCNC NT65-3X card is poor in terms of filtering capacitors (cost reason, I guess).

Cheers,

Rak

[helisat]

Hi Rak,

You are correct, the board does lack the required filtering. Three capacitors, it's a shame that YOOCNC have overlooked this.

Many thanks for your help. My machine is now running very smoothly, and no longer skipping steps.

Cheers,
Luke

[LeslieTroyer]

Color me slightly confused -- Looking at the drawings - I see a 47 micro-F cap in parallel with the 1 micro-F cap -- but 470 nano-F is .47 micro-F or 470K Pico-F. I did install a 470K pF cap and it seems to be better - but wondering if I'm several orders of magnitude off on the filering and will it cause problems??

Thanks
Les
BTW - my velosity is 100 and 25 for accel (inch base units) I think I can push faster - but this is good for now.

[Rakounet]

Hi Leslie,
At low frequency, the filtering is done by the 47 micro-F cap, and at higher frequency by the 1 micro-F cap. If by adding a 470K pico-F cap, you see an improvement, may be it is because the 1 micro-F cap does not play its role at higher frequency. It depends on the quality of the cap. That why, for this kind of application, ceramic caps are better at higher frequency.
I don't know if it answers your question. Let me know.
Rak

[helisat]

Hi Leslie,

Sorry for the late reply. How did you go with your capacitors? I followed Rak's advice, and put 470nF between the GND (directly on the Toshiba IC pin) to the Supply Rail. Please see the photo that Rak posted below. You repeat this for all 3x Toshiba IC's.

I was unable to find a 470nF in a ceramic, so I used a 470nF MKT capacitor. Adding the cap completely resolved the skipping steps, and the movement 'sounds' smoother.

Hope that helps.

Cheers,
Luke

[LeslieTroyer]

I'll get some 47 micro-F today and see if it sounds better. the 470K Pico-F seem to stop the occasional dropped step I got when going -X, but the drives now have a high-freq whine to them that wasn't there before. I'll probably do the same to the 4th axis which is on a seperate board...

Les