Laser PSU noise, EMI and grounding

[Snin]

Hey all,

we've been running a 80W laser cutter with the "old" lightobject dsp for over a year and recently switched to the new awc 608 DSP. During this period we've had several times vague problems. I am now pretty sure these problems are caused by electromagnetic interference caused by the laser power supply. I'm wondering if anyone else has had problems like this or knows good solutions.

In more detail, machine setup:
Our machine is driven by Whale3 servo drives, controlled with step and direction signals from the DSP. The rest of the electronics is fairly straightforward. All The electronics are built into a steel enclosure with separated compartiments for the DSP, laser PSU, servo drives and other power supplies. All the digital grounds are connected in a star scheme to one point. The main enclosure is connected to the mains-earth.

Problems:
When we run a complex cutting job with a lot of start- stop movement (for example a lot of small holes) or during raster engraving, the zero position of the laser head after the cutting job deviates from the zero position before cutting. The deviation is in the order of 2 to 10mm. Some days it is on the x-axis, some days on the y-axis, some times there is no deviation at all.

It is motor problem as the servo drives have encoder feedback and do not show a position error after the job. Backlash is also not cause this big deviations.

Now here is the thing: if we run the same job with the laser off (for example with the lid of the machine open). There is no deviation at all, never. We've tested this dozens of times.

My guess is that the laser power supply is somehow distorting the communication between dsp and servo drives, or between servo drive and encoder. I've been measuring with my oscilloscope on the step/dir lines and indeed, I can see some spikes of around 2V on these lines when the laser fires. Not good.

So my question: how to avoid this? the dsp and servo drives are already in different compartiments in the enclosure. Can it have something to do with proper grounding? Should digital ground and enclosure shield and mains-ground be connected? We've already put some ferrite beads on a lot of wires but it doesn't seem to make any difference.

thanks in advance,
J

[djsuthar]

Dear Sir

I have face this problem in our machine. suggestion is mention us under.

1. keep big distance between Laser power supply and controller. it mean don't fix up power supply near to motion control card.
2.use only shield cable.
3. Add bus bar in the control panel of machine.

[Tech_Marco]

Use a EMI noise filter for the laser power supply and the DC power supply. Then, add one or two big capacitor to the DC power supply output terminals where a DSP is connected over. The Cap should be running at 50V 5,000uF or 10,000uF

http://www.lightobject.com/Search.aspx?k=emi

By the way, make sure to ground everything, I mean "everything": Laser power supply, DC power supply, Drivers, DSP control mainboard, laser machine case.... And, make sure the ground to the AC outlet is 'true ground' and well connected. Some machine coming from China with the Ground unconnected "Fake grounding"


Marco

[Snin]

Thanks for the response. I think we did all of the stuff you mention.

Our electronics are set up to prevent interference, see the attached image. Laser power supply is right top, DSP is left, servo drivers are right down, 230V wiring and psu for dsp and psu for motors are in the middle. All data- and sensor cables are shielded cable (SFTP cable). The mains 'earth' is connected to the big enclosure, to the laser psu enclosure and all other metal casings by means of the star earth point that you can see on the bottom of the enclosure. All cable shields are also connected to this 'earth net'.

The motors are fed from their own 70V power supply. This supply is not connected to the 'earth', only to the drivers. The manual explicitly states that it should be wired like this.

The ground from the psu for the dsp is at the moment connected to the 'earth' only through the laser PSU, as control-input-ground and the enclosure are connected internally in this device. We have not made any other connections between digital ground and 'earth' because it would form a ground loop.

There is an EMI filter at the mains input. We've also added a 10.000uF capacitor over the dsp power supply. Maybe we should add another EMI filter just before the laser PSU>

Check the two scope images. The one with the big spike is what we get when we measure between digital ground and the X (orange) and Y (bleu) step lines and we fire the laser. This spike only occurs when the laser starts to fire, for the rest of the time the line is clean from noise. This is measured with the probes on 10x attenuation. We measure the same kind of spike on the digital power line and almost any digital line.

We've measured the same spike with the laser psu plugged into a different power group with no other electrical connection to the machine at all, not even the shield (we fired it manually).

We're puzzled how the noise comes into the digital power supply and other lines.

The only thing we came up with is try to filter away the noise with ferrite beads and low-pass filters on the step lines. We were able to suppress the spike a fair amount using a 500ohm resistor in series with the step line and a 100n capacitor to ground behind it. Problem is that this filter also distorts the step pulses. The problem is less apparent now but it's still there. (see other scope pic) Furthermore it seems to differ each day how much distortion we have on the machine position. I'll try and fine tune the filters later and let you know how it works out.

[Tech_Marco]

How did you hook up and setup the "HV" cable? From the picture you posted, it look like that the HV cable was extended and tap with some sort of electric tape, isn't it? Can you shoot me a closed-up picture of your HV route? Another suggestion is to install an insulation tube to wrap up the HV cable like the silicon tube that selling on my site. If you can get some in your local, do so. Check the voltage appear on the laser power supply in respect to the ground see if any voltage build up. If it does, it means that it is "Ground" issue. Or if you can, swap another power supply to isolate the problem.

Marco

[Raman Narayan]

From all the info. I see the issue to be due to laser power supply switching getting coupled into the
power supply. I would like to add that if you use a low ESR capacitor, such as 100uf tantalum capacitor
(with voltage rating 2-3X the dsp controllers ps voltage), and then connect it directly to the power
supply input terminals at the DSP, it should take out most of the coupled glitches from the power supply.
The large cap on the powersupply is useful only for lower frequency noise suppression, since the
caps do not have a low ESR.
You could also try some of the exotic mu metal shielding plates as a divider
between the laser power suppy and your low voltage supply for even better emi isolation.

[Snin]

Dear all,

it's been a while (we rushed back into work when we solved the problem) but i would like to share the results. Marco was totally right that the problem was in the HV cable routing. It turned out the HV cable was running very close (1mm) to the aluminium frame of the machine (not visible on the picture). The HV switching was coupled into the frame, which in turn is grounded. The solution is very easy: support the HV cable by non-conduction material so that it doesn't come closer than 50mm from any grounded metal parts. Silicone tube gives some space indeed, but I suggest to keep an even bigger distance just to be safe.

Thanks for the help!

[medicdude]

HV cable was acting as a capacitor to your grounded frame, which is connected to your stepper motor ground. Increase distance from + HV cable to - HV cable, from each cable to case, and add shielding to the HV cable to reduce capacitive coupling.

Insulate your stepper motors (do not ground your stepper motors, use plastic washers/screws to connect them to metal case). +15V does not need to be grounded and it's a good idea not to ground it so you don't get interference through the ground. Obviously the motor PSU itself should be grounded as it's plugged into your 110V/220V.

Your motor DC power supply itself *should* provide filtering at the output, you can increase this by adding caps between the +- motor rails. AC line filtering on the primary side of the power supply isn't necessary unless you're actually using AC to drive an analog signal. You don't need a bleeder resistor but a low-ohm in-line resistor is a good idea downstream of the additional caps to safeguard shorts between the rails.

You don't want ANY capacitance between your HV rails, this will just increase the time it takes to get to voltage, increasing the draw of the laser supply and interfering more with the other supplies on the line. If you want to add capacitance here you need to do it in the power supply itself, and you'd need to take a look at the schematic to see where to add it.

Do you have any schematics on the HV power supplies Marco?

The problem could be with the chopper inductors in the HV PSU inducing a voltage on nearby lines, you'd want to shield the HV psu in some type of faraday cage to minimize this, a box of cold rolled steel would work well, that thin aluminum case probably doesn't intercept enough of the magnetism.