Troubleshooting possible PSU induced EMI

I am troubleshooting lost steps on a cheap CNC conversion.
My setup is:
I use a set of steppers and PSU from a known seller of kits for Seig X2 mill conversions.
I also use a Gecko G540 driver and UCCNC UC100 motion controller. The G540 is is bolted to the back plane but no grounding wire to the back plane.
The PSU is grounded to the back plane and bolted to the back plane. The mains are grounded to the back plane.
The above are in a 20"x 15" x 10" steel enclosure.

I am trying to track down my lost steps, which I know can be a rabbit hole, so as part of the process I have a couple of questions:
- If the G540 and UC100 controller were within about 1.5" to 2" of the PSU, could that be considered too close and a source of problems?
(NOTE: I have moved if the G540 and UC100 about 6" apart and am now testing)

- Would it help to build some sort of Faraday cage around the PSU? Using metal window screening?

Quote:

---

Originally Posted by cncuser1

I am troubleshooting lost steps on a cheap CNC conversion.
My setup is:
I use a set of steppers and PSU from a known seller of kits for Seig X2 mill conversions.
I also use a Gecko G540 driver and UCCNC UC100 motion controller. The G540 is is bolted to the back plane but no grounding wire to the back plane.
The PSU is grounded to the back plane and bolted to the back plane. The mains are grounded to the back plane.
The above are in a 20"x 15" x 10" steel enclosure.

I am trying to track down my lost steps, which I know can be a rabbit hole, so as part of the process I have a couple of questions:
- If the G540 and UC100 controller were within about 1.5" to 2" of the PSU, could that be considered too close and a source of problems?
(NOTE: I have moved if the G540 and UC100 about 6" apart and am now testing)

- Would it help to build some sort of Faraday cage around the PSU? Using metal window screening?

---

First off. Is the UC100 a genuine one from cnc drive? There are fakes around.
Secondly which stepper motors are they?
What voltage psu and does it / they have the wattage to run all together at same time?

My psu's (*2), drivers & board are very close together with no issues on my X2 conversion.
However I am only using 36vdc and it isn't enough for my 425oz nema23 steppers.
I work in mm.
If I set acceleration above 120 it misses steps and can stall.
If I go above rapid velocity of 1300mm/min I get the same thing.
I should be running at least 48vdc because inductance is 3.5mh. This would increase performance.

I now have a bigger machine (PM25 size double the weight) with 566oz nema24 steppers 3mh inductance.
This time I am using 60vdc and my psu's, drivers, board again are tightly packed in the control box.
I can accelerate at 500+ and run a rapid velocity of 3000mm/min+ without any problem.

The higher voltage has bought everything to life.
Furthermore. My cables are not screened but I have no physical limit / home switches yet.

dazp: I bought a software licence from UCCNC after sending them my UC100 info, so I assume the UC100 is original.
The PSU is 48V/7.3A
The steppers are 270 Oz/In.(NEMA 23), inductance of 2.8mH.
I was running at 12 In/min with accel at .5in/min^2, which I think are very slow for an X2.
I also had this problem cutting air, no forces at the tool ( but of course I first saw the problem in my work)

When the machine was running I could push against axis movement with almost all my strength and the motors would not stall!

I am not an expert so maybe there is a possibility that the PSU is not working properly or not powerful enough? How could I verify this?

Quote:

---

Originally Posted by cncuser1

dazp: I bought a software licence from UCCNC after sending them my UC100 info, so I assume the UC100 is original.
The PSU is 48V/7.3A
The steppers are 270 Oz/In.(NEMA 23), inductance of 2.8mH.
I was running at 12 In/min with accel at .5in/min^2, which I think are very slow for an X2.
I also had this problem cutting air, no forces at the tool ( but of course I first saw the problem in my work)

When the machine was running I could push against axis movement with almost all my strength and the motors would not stall!

I am not an expert so maybe there is a possibility that the PSU is not working properly or not powerful enough? How could I verify this?

---

Just add an EMI Power Filter if you think there is any noise causing your problem UC100 do have problems with noise, add some ferrite cores at each end of the USB cable if you don't have it will also help do you have a VFD Drive running your spindle this too needs a EMI Power Filter close to the input power supply to the VFD Drive

Mactec: re the EMI power filter, that goes on the mains line into the project box? I could add it, but my suspicion was the the control electronics was getting EMI from being too close to the PSU. Maybe I am wrong but I thought the EMI went through the air between the two units ( radio wave EMI?) I didn't think the EMI went through the wiring from one unit to the other.

But now you made me think maybe the wiring from the G540 to the steppers is too close to the USB cable for the UC100. I also had trouble with losing the USB connection. I thought I took care in using shielded and grounded wires for the steppers and avoiding ground loops.

Hello.

One question. Does you lost steps problem appears in all the axes of your machine?

In the last few days two guys showed up with the problem of lost steps but only in one axis. One already confirmed that the origin of his problem was some mechanical mismatch. While the second has not posted what has happend this far from his comments I suspect his problem is again of mechanical origin.

Did you ever had the machine running properly before or are you just finishing the conversion?

Would you be willing to substitute at least one power supply with a linear one? This comment comes from the fact that the machine of one of my customers got his SMPS damaged and we replaced it with two linear PSs. The drivers used are DM860 6A units.

Linear Power Supply. To me it´s a fancy name for what we made. A simple 5 Amp 120 to 24V transformer wired as a voltage doubler effectively delivering about 62V when loaded. 4 6A2 diodes in two pairs of two paralleled diodes each (no other diode was available at the time) and 4 15,000uF/50V electrolytic capacitors wired as one 15,000uF/100V cap. Two such power supplies were assembled to replace the original SMPS and they have been working fine for nearly a year. While the owner bought a new SMPS he decided on keeping the linears.

Note: a comment on my part. In the particular case of my customer I suspect that the SMPS failed due to voltage spikes generated at motor direction change. He is manufacturing pillows with a zig-zag pattern where the sewn vertices have to be perfect and the stitches per inch have to be also perfectly constant. Thus the movement change has to be sudden effectively generating spikes being sent back into the power supply. While i had not had that problem before I see no other reason for the SMPS failure. The linear ones by not having very sensitive components and the electrolytic caps at the output can easily absorb those spikes.

The included diagram shows the correct circuit although the output voltage is incorrect. Also please remember that I mentioned that two diodes in parallel are being used instead of a single one. There should be single diodes with the required voltage and current capacity and could be used in any actual PS. Also adding a 1.5Kohm to 3.3Kohm 5W bleeder resistor connected at the output would prove very convenient.

Let me know what happens.

Regards.

Why not use a transformer with a suitable secondary without going to the need for those huge capacitors, also much more efficient when the correct secondary is used.
For servo's and steppers I only use linear supplies, never SMPS.
Also I always earth ground all DC commons, no problems with EMI so far.
Al.

Quote:

---

Originally Posted by cncuser1

Mactec: re the EMI power filter, that goes on the mains line into the project box? I could add it, but my suspicion was the the control electronics was getting EMI from being too close to the PSU. Maybe I am wrong but I thought the EMI went through the air between the two units ( radio wave EMI?) I didn't think the EMI went through the wiring from one unit to the other.

But now you made me think maybe the wiring from the G540 to the steppers is too close to the USB cable for the UC100. I also had trouble with losing the USB connection. I thought I took care in using shielded and grounded wires for the steppers and avoiding ground loops.

---

No that is not how it works each suspect device needs to have there own EMI Power Filter mounted close to there input power supply connection for that device

If you did not Ground the shields correctly then the shield will not do it's job it is very hard to create a Ground loop in a system that is only using ( 1 ) mains Power source

You can also shield the electronics with a simple aluminum cover

mactec: Does EMI Power Filter get connected DC input of the Gecko G540 or only mains voltage input of the PSU?

"You can also shield the electronics with a simple aluminum cover" For the cover, are you referring to PSU or control electronics?

BBMNET: I have a simple arrangement. One linear PSU, one Gecko G540 with a UC100 attached. Everything inside a steel box. Inside the box the stepper wires are shielded and grounded to the G540 only. The PSU is 6" away from the G540 and UC100.

You asked about a mechanical origin. The machine is a cheap Chinese mill, Sieg X2. Mechanical sources could be slipping balls inside the ballnut or motors losing steps. But I have never heard of balls in ballnuts slipping and as i mentioned, with most of my strength I can't stop the axis from moving.

The amount of steps lost is about .0015" on each axis after about 340" of travel, using about 11K lines of code, 3.5 hours of travel time. This is on 2 axis, the third axis is not measured for loss.

Quote:

---

Originally Posted by Al_The_Man

Why not use a transformer with a suitable secondary without going to the need for those huge capacitors, also much more efficient when the correct secondary is used.
For servo's and steppers I only use linear supplies, never SMPS.
Also I always earth ground all DC commons, no problems with EMI so far.
Al.

---

Hello.

I must say that I do agree with you 100%. However in mexican provinces it is not precisely easy to find the parts that you may need. In many cases you have to buy them from Mexico City with all the delay that such a thing implies. That is of course more noticeable during weekends.

This particular customer had the need to complete one order to be delivered first time in the following Monday so we used what we could get. Fortunately we found parts for what we needed in a tianguis (some kind of street market where you can get all kinds of used stuff) but at the time even the idea of using car batteries crossed our minds.

Right now the only thing I would change in the power supplies are the capacitors. Having 8 caps with connections too close to one another is one thing I don´t like. Recently I added silicon glue as the owner didn´t allow me to change the caps. He doesn´t like to see his machine stopped. He knows the risk but well, he is the money man.

Thanks for your comment.

Regards.

You can watch a video of the machine I was referring to at the following link. Notice the SMPS (48V/12A).

https://www.youtube.com/watch?v=ErNjNtC60As

I should mention that small familiar shops are very common in Mexico. They usually live out of one single machine and are usually pressed to deliver on time or risk loosing orders. Cancelation of delayed orders is very common where the buyer later offers to buy their delayed production at an average of 50% of their value.No partial payments is given to the shops so they have to risk their money buying the fabrics, threads and whatever they need to accomplish their orders. Thus the need to deliver on time.

Quote:

---

Originally Posted by cncuser1

mactec: Does EMI Power Filter get connected DC input of the Gecko G540 or only mains voltage input of the PSU?

"You can also shield the electronics with a simple aluminum cover" For the cover, are you referring to PSU or control electronics?

---

Control Electronics

Mains Voltage only Ac

The EMI Power Filters you would only use on the Ac inputs to the main power supply that powers the device you want to reduce the EMI, they need to be mounted close to the device

Quote:

---

Originally Posted by cncuser1

BBMNET: I have a simple arrangement. One linear PSU, one Gecko G540 with a UC100 attached. Everything inside a steel box. Inside the box the stepper wires are shielded and grounded to the G540 only. The PSU is 6" away from the G540 and UC100.

You asked about a mechanical origin. The machine is a cheap Chinese mill, Sieg X2. Mechanical sources could be slipping balls inside the ballnut or motors losing steps. But I have never heard of balls in ballnuts slipping and as i mentioned, with most of my strength I can't stop the axis from moving.

The amount of steps lost is about .0015" on each axis after about 340" of travel, using about 11K lines of code, 3.5 hours of travel time. This is on 2 axis, the third axis is not measured for loss.

---

Hello.

The slippage I was referring to usually happens at the couplings. Yet I have seen in more than 40 years of professional work only one machine where the ballscrew had to be remachined and the balls changed due to excesive wear. The following link shows one machine we retrofitted years ago. If you watch carefully at the end of the video you will barely notice a twin machine behind the first. That second machine is the one where the ballscrew had to be repaired. Then we retrofitted it too.

https://www.youtube.com/watch?v=SC3OrgzDfJU

Please try noticing the wiring of that machine. It is not particularly neat. In fact it is a mess. Yet we had no functional problem. We had to wait for some time to properly arrange the wiring.

Just to make sure I´m on the right track, what are your microstep settings and the tolerance of your system? Do you have information on the accuracy and repeatability? Due your numbers it is possible that you have set a too high tolerance or asking for something your accuracy and/or repeatability can´t give. By the way, please confirm if your machine ever worked fine or you are retrofitting it.

Regards.

PS: may I again suggest that you disable the feedback if possible? That way you could confirm where the problem can be or at least get an idea. Theoretically running open loop should give you the same results unless some obstruction or problem gets the system to loose steps. The closed loop system checks for a position match and compensates accordingly. A function an open loop system cannot perform.

I Would fit the EMI filter regardless. They will be mandatory on any more machines I make. They were the biggest noise reduction step I did.

rodw: Can you help me understand something, maybe I am mistaken. I assumed that a PSU could generate EMI due to it's internal circuitry, my suspicion was this was causing problems. Would an EMI filter address that?
And the benefit of the EMI filter is that it would filter out EMI from the mains? Does this EMI pass into the electric components through wiring or from radiowaves emitted from the mains wires?

Quote:

---

Originally Posted by cncuser1

rodw: Can you help me understand something, maybe I am mistaken. I assumed that a PSU could generate EMI due to it's internal circuitry, my suspicion was this was causing problems. Would an EMI filter address that?
And the benefit of the EMI filter is that it would filter out EMI from the mains? Does this EMI pass into the electric components through wiring or from radiowaves emitted from the mains wires?

---

I'm no expert but noise is a funny thing. It might be a bit different for you as I was working on a plasma cutter and I was getting a noisy arc voltage signal. I bought a cheap USB oscilloscope to investigate. I concluded the bulk of my noise was on the mains power supply (you can still see an AC ripple in the DC supply). So I added an EMI filter (just replaced the IEC connector with a filtered one). Progressively, I removed some of the work arounds I had added. The only other thing I have done is add separate earth wires from every motor case back to a star ground point so there is a path over any linear rails.

So by far the biggest improvement came from the IEC filter. For $10 its cheap insurance!

Quote:

---

Originally Posted by cncuser1

rodw: Can you help me understand something, maybe I am mistaken. I assumed that a PSU could generate EMI due to it's internal circuitry, my suspicion was this was causing problems. Would an EMI filter address that?
And the benefit of the EMI filter is that it would filter out EMI from the mains? Does this EMI pass into the electric components through wiring or from radiowaves emitted from the mains wires?

---

Yes through wiring is always the biggest problem for noise, and how wiring is done, Twist all DC wire pairs where ever you can the same for Ac wires don't Twist the Ground wires in with the Ac Power wires

BBMNet: This is a manual machine I converted to CNC. It is a small cheap Chinese mill. Sieg X2, with 270OZ-In steppers and ball screws. it is a simple machine.