Grounding and RFI

[asteude]

I know this topic has been beat to death, but I'm still not exactly certain what to do.

I have a PRSstandard with a 2 HP Spindle (single phase 220V) and a Yaskawa V74x Drive. Whenever I run the spindle, it knocks out the DSL internet and TV (AT&T Uverse) in our home. My wife is very patient with me, but it's starting to really irritate me.

I contacted the Yaskawa service shop in our area (North Carolina), and the engineer was very adamant that I should make sure my grounding was up to snuff. After reading all sorts of posts on related issues on the ShopBot forums
(a) I am really wondering why in the world ShopBot doesn't provide grounding kits and instructions for their machines!?
(b) I am convinced that my machine is not grounded properly
(c) I am still not entirely sure I know how to ground it properly.

Since the parts of the table are all powder coated and nicely insulated from each other and the spindle certainly isn't reliably connected even just through to the X-axis rails, I've read that each of these parts have to be electrically connected before grounding the whole mess to the same earth ground as the control box (which in my house is the neutral bar in the breaker box.)

So, I'm wondering what sort of materials and methods other have used to ground their ShopBot. Don't I really just need to ground the spindle? And if so, what sort of connector and connection should I get and where and how should I fasten it? I'm a tad hesitant to just drill a self tapping screw into the spindle housing. If I need to ground the rest of the components, the same question goes for them. What type of connector? How did you connect it and where? And finally, on the other end, do you just slip the ground wire under the cover of the electrical outlet and connect it to ground there?

The machine has performed amazingly well considering that, on closer analysis, I haven't really grounded it at all.

I'm looking forward to seeing your solutions and I extend my sincere gratitude in advance.

Andy
www.RockingFrog.com (http://www.RockingFrog.com)

[beacon14]

I looked at each sub-assembly as a unit, with the idea of not expecting the rollers-to-rails connection to be a reliable ground. That means I grounded the main table, Y gantry, and Z axis individually, all to the same ground location, which in my case is the electrical conduit nearest to the ShopBot. I know the conduit is well-grounded because I installed it myself, and every junction box has a ground wire screwed to the box which runs back to the main panel.

To attach each ground wire to the machine I used a grinder or a file to remove a little paint and bolted the ground wire onto the bare metal. I also removed one bolt at each connection point (leg to main rail, the corners of the Y gantry, etc.), ground off a little paint, and re-tightened the bolt.

I made sure the dust collector and hose was well-grounded, but did not connect the dust collection hose ground to the SB, only back to the dust collector frame. If any item is grounded to more than one location you can get a "ground loop" which can be as bad or worse than no ground.

[Gary Campbell]

Andy...
This may give you a few ideas: http://shopbotwiki.com/index.php?title=Grounding_your_ShopBot

[Brady Watson]

There is absolutely no way to stop radio interference from the CNC without some kind of cage to block the RF signals.

You need a faraday cage: Cheap Trick (http://www.uline.com/BL_54/Static-Shielding-Rolls)

-B

[dana_swift]

Grounding has little value in keeping signals from radiating OUT from a system, or preventing reception of outside signals that interfere with the digital data our bots use.

Brady is correct, a Faraday cage is needed. Shielding works, because it is a faraday cage wrapped directly around the offensive signals.

As a side note- any component, such as the spindle VFD that creates interference has something very wrong. No surprise the manufacturer points the finger at some other source. They must have a big problem and not interested in fixing it unless it interferes with their sales.

I don't have any elaborate grounding system, and have no problems with generating interference or unreliable operation.

My suggestion, take an old transistor radio, tune it between stations, and hold it up to various cables and control boxes and see which one is causing the most noise on the radio. Find the source. If it is cables, you can enclose them in screen wire, but I suspect the motors or controller boxes. They used to be made of metal, now often plastic.. hence the suspicion.

Good luck-

D

[frank]

Hi Andy,

The RFI problem is pretty rare and the 2 or 3 cases that I have seen were solved with an RFI filter at the power input. The spindle manuals show these optional filters. If you would like to order one of these filters, give me a call.

[chiloquinruss]

In addition to being a botter I'm also a ham radio operator and in the case of rfi interance there are always two sides to every story. The spindle is suspected of being the transmitter but the cable connection is the other side of the story. By law the cable company boxes are supposed to be shielded. Most are not. If your tv, internet etc, is all connected to the cable company you may want to check out their connections as well. If its a little plastic box, try wrapping the box with a little piece of aluminum foil and see if it helps at all. If it does, contact the cable company and explain the problems. If they can't or won't help drop a line to your local FCC and they will get er' done! :) Good luck. Russ

[srwtlc]

Andy,

I had the same problem where mine would knock out my DSL. After installing the RFI filter as Frank mentions, the problem has ceased.

Around here (probably most places), the phone company shares grounds with the power company.

Scott

[asteude]

All - Thanks for the replies!

Everyone picked up on my mention of RFI and keyed in on that pretty quickly. I probably shouldn't have mentioned it because I really wanted to address the grounding issue (which I'm sure exists regardless of the possibility of RFI.) I also probably should have used the term "line noise" instead of RFI. I believe that's what the "RFI filters" actually address anyway. The RFI goes away as a side effect of reducing the line noise on the powerline. But it's the line noise that (IMHO) kills the DSL. I had already tried a faraday cage around the drive as well as the DSL modem. I even started "shielding" my ceiling with aluminum foil! No space aliens were going to tap into my controlbox fer shore! :o Anyway, none of that helped. But, before I buy the filter(s), I'm going to fix my grounding. It needs to be done and if I'm lucky, it'll take care of my DSL issues too.

I just read Gary's ShopBotWiki page on Grounding (http://shopbotwiki.com/index.php?title=Grounding_your_ShopBot)and it's exactly the sort of detail I was looking for! THANK YOU! It should be required reading for every ShopBotter. It also made me re-think my theory about reducing line noise by grounding the machine. Perhaps the two are completely unrelated and I'll be buying that filter shortly after all.

Thanks for everyone's input. I'm off to Home Depot to buy a spool of green wire and assorted nuts and bolts.

Andy

[asteude]

Alright, I guess I'm drinking the cool aid. The grounding didn't help at all. Now it's just a matter of figuring out why ShopBot is selling a different RFI Noise Filter (from www.eecontrols.com (http://www.eecontrols.com)) than the one Yaskawa specifically makes for the drive. Very odd.

[dana_swift]

One thing you can do is use twisted pairs for your wires. Each twist of the wires will radiate the outbound energy 180 degrees out of phase and tend to cancel the radiation at long distances.

Back in the old days of land lines, the nations phone system was largely unshielded, depending on the effect of twisting the wires to keep EMI out and crosstalk in. Much less expensive also.

I still suggest my transistor radio sniffer. Find where your problem is and fix the actual source, it has often occurred in my experience to be something very different than what my first guess was.

As to the filters being different, that is not surprising at all. What is needed is a low pass filter that will let 60hz through but block the kilohertz and up noise. The exact cutoff frequency hardly matters, and almost all of the filters you can buy are a single stage Pi (CLC) filter. Effectively they are all the same, the amount of attenuation you get at RF frequencies will be almost identical. Buy a unit based on their current carrying ability, and price.

D

[asteude]

Hey, Dana - I'm familiar with twisted pairs. This girl I used to date had a twin sister and... OK, never mind that. I've never heard of twisting to eliminate outward radiation of electromagnetic noise. But, I don't know why it shouldn't work in that direction. The only problem would be to twist that half inch 220v power cable. Yikes! I'm not sure the insulation would survive it.

More significantly, I think we have to distinguish between the electrical noise introduced by the spindle and carried on the power line and the electromagnetic fields which radiate outward from the cable. I have a hunch that the more significant problem is the noise carried on the power line to the circuit box and, from there, distributed to the rest of the house. While shielding and twisting may be helpful, it only addresses the symptom.

To address the line noise, I'm getting passionate advise in two different directions. The Yaskawa engineer insists on improved grounding. We already know that grounding the ShopBot didn't help. And I'm not sure how I can significantly improve grounding for the VFD. In the other direction, I'm told that a filter is the way to go.

The Yaskawa manual talks about two different input side filters which apparently serve a different purpose. So far, the Yaskawa engineer has been steering me away from the filters and has avoided explaining the difference. There's some pretty sparce documentation for one of the Yaskawa filters which explains the wye/delta capacitor configuration used (http://www.yaskawa.com/site/dmdrive.nsf/link2/MNEN-5JFUKD/$file/2Y25-0349.pdf) to achieve better RF attentuation than a wye configuration. The eecontrols filter sold by ShopBot is a double stage filter. The spec (http://www.eecontrols.com/documents/Detaslinefilters.pdf) unfortunately only goes into detail on the line of single stage filters. And if you know what any of this means with respect to the problem at hand, I would love to get your input.

At the moment I'm still frustrated by even the so-called experts' apparent inability to give me useful information. If I find out more I'll post here.

Andy

[dana_swift]

Andy- the power cable I used on my bot is Carol 12/4 (a General Cable product), my vague recollection is this is internally twisted as a group. There are multiple 12/4 products from GC, I can get you a specific part number if you like.

That rubber sheathed cable goes from the SB to the junction box, after that the power lines are enclosed in metal conduit.

One other thought about your emi problem. Be sure all of your connections are tight. (Power off before checking!) Nothing emits broadband RF like arcing, and its a huge fire starting opportunity I'm sure you are aware. One thing to do is just put your hand on the power cable near junctions and plugs to see if it is heating at all. Heating is a sign of trouble. The cables should be very near room temperature.

---


Oh.. feel free to tell more about those twins. Its of no interest to the general community I am sure, but I would be happy to study the situation carefully for its technical merits ;)

[dana_swift]

Footnote- just looked at the info from Yaskawa, that is a "C only" network, which I would not recommend. The ones from EEC are "CLC" Pi filters like I spoke of before, they would be much more effective.

THe EEC products are marked as single phase, if your spindle has single phase power its just a matter of getting TWO of the D-MSC 30 (for the current requirements [assuming 30A is enough for the spindle]). If you have a three phase power requirement, then look around for a three phase filter equivalent to the D-MSC 30.

[asteude]

I have the 30Amp single phase spindle. Scott Worden installed the D-MDC 30 which, annoyingly, is not described in detail in the PDF (only superficially on the first page.) This is what's sold by ShopBot and seemed to work for him. It's the double stage version of the MSC filter.

I think you just convinced me to go with the eec filter though.

Thanks so much for your input! I really appreciate everyone taking an interest.

Andy

[asteude]

By the way, I just found my old transistor radio (for you youngsters - that's an iPod that streams music wirelessly for free) and tried running the spindle. The results were pretty spectacular: as soon as the spindle started running (at .5 or .6 on the display), there was significant noise on the AM band. I didn't bother with FM since AM is closer to the DSL frequency range than FM. As long as I was near the Shopbot, there was a lot of noise. It definitely got worse near the spindle and the cables. And it didn't seem to matter at all where along the power cables I checked. So, there was significant noise from the spindle all the way back to the wall outlet.

[dana_swift]

Now you have found the source of the emissions. Using the suggestion of a pre-ipod music streaming device (I love that description). You know what specific problem you are trying to solve.

That technique has worked for me before. I learned about it from an RF engineer I used to work with. He taught me all kinds of practical things in working with RF I have never seen elsewhere. Glad I could pass it along.

The EMI filter should help, and twisting should help again. Between the two it should make a significant difference. Even then the interference will still be there, but it should be much less. Metal rigid conduit or flex conduit should provide a solution also (as it is a faraday cage).

[Slim]

Could not having your shopbot grounded well enough lead to crashes? lately i've been running just fine and then i get communication error. I'm gonna ground it better and see if that fixes the problem.

[asteude]

Tim -
It sounds like a lot of people have fixed their communication problems by properly grounding the bot. See the excellent ShopBot Wiki page on grounding: http://shopbotwiki.com/index.php?title=Grounding_your_ShopBot

Also, make sure the thin control wires are as far as possible from the power cable to the spindle. As discussed in this thread, the spindle feeds a ton of electrical noise back into the powerline. The noise in the line translates into electromagnetic fields which are much stronger closest to the power line. If your control wires are placed too close to the power line, the electromagnetic fields will induce fairly strong electrical currents (noise) in them. And, of course, you don't want electrical noise competing with the commands you send along the control wires.

All -

I was able to find the EE Controls' (www.eecontrols.com (http://www.eecontrols.com)) D-MDC-30 Power Line Filter at their NC distributor, www.controlsource.com (http://www.controlsource.com), for $95. With a little luck, in a week or so I'll be surfing ShopBot forums and DVR'ing my wife's TV shows while my bot makes sawdust.

Andy

[asteude]

I received my EE Controls D-MDC 30 Powerline Filter ($105 including shipping from Control Source in NC) yesterday and installed it (in a separate box outside the VFD.) So far so good! I currently have the spindle running at 18,000 and have no degradation in DSL.

Thanks for all your help, ShopBotters!

Andy

[dana_swift]

Congratulations! I like to hear that a problem was solved, and how.

:)

D