EMI from turning overhead fluorescent light On/Off kills print

I still have an EMI problem with my old Gen3 electronics.

Whenever I switch the rooms overhead fluorescent light On/Off EMI kills the print.

Obviously either the Motherboard or USB is being affected.
I have a ground wire between the Motherboard and Extruder Controller and I have ferrite cores on my USB cable.

What else can I do other then enclose the electronics in a Faraday cage (which I don't want to do for various reasons)?

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

... make the ground connections between the components and to the PS shortest possible and best connected to one point near the PS.

A good idea is to connect ground of the PS and your computer housing too - measure prior, if there are big differences.

And use the same line for all all connected parts ... especially PC-PSses can make curious things with ground levels ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I am using a simple laptop regulated 12v power supply (6A).
My heated bed runs off 220-240 vac using a 12v control / 220-240 vac load relay from the FET.

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

... sometimes it's as simple as rearrange your cables - try to avoid loops in the source and data wires and best add a grounded shield around your data cables ... and don't feed your 'logic level' wires near the AC wiires ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

@Viktor: Would aluminum foil work as a test to see if the shielding fixes the problem?

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

... should work, when grounded ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I have the similar issue with gen 3. I have also fluorescent light, but thats ok with me. It can be with drosser (electronic "ballast"), or with classical transformer solution, mines are with electronic solution (dunno if these are right english words but should sound smth like that).

What does kill my prints in my setup is starting of 800w motor of a disc cutter in another room. I rarely use that motor so i easily settled with not using it at same time. But my initial guess is that its something along the effects of reactive power use, which probably propagates along the mains (even maybe along the null) and goes beyond what normal psu can mitigate, probably affecting comms first.

So a good question for me, btw, does it happen same thing with sdcard prints? If that can be tested that is.

In the case you have a classical solution for fluorescent lights, e.g. transformers, can try use electronic solution instead (here generic name is "drosser" - which sounds like english but dunno if its actual an english word). Also what i had in mind at first, can try use one or two "motor starting capacitor" before neons which might help if big enough. Normal reactive power capacitors are rated in kVAR, but i doubt there are some that small as for a normal household, those are for industrial use. Maybe there is some smaller solution for main-fuse-household-panel-power-entrance (again dunno right words). Coz I also guess this whole thing probably doesnt happen in normal condition but it is probably agravating in a context where the cosPHI is already bad, say more reactive power consumers than average and further from the medium-to-low V lines transformer, so in something like this the cosphi from the start is already much lower than (here) legal ~0.92 (i think the 0.92 is averagelly accepted due to max distance from power line transformer to a certain use point).

Its interesting that also in my case, it doesnt happen when the fridge or washing machine starts. I guess these have emi filters built in that may help a little if the cx cy and inductors values are high enough. Maybe adding a emi filter like ones rated for welgind tool could help. Or if you have an old broken tv or pc monitor you can find the exact emi filter parts at the power entrance and can DIY instead of buying. Atx psu usually have emi filter space (provisional) on the board, but that is usually empty, or at least the inductors there are missing, so usually those only have a 1uF cap and some cy and thats it. Emi filter parts missing in atx psu are quite a pitty considering its a smps and not just any power user.

Again obviously everything i said are guesses, so pls treat them as such. And if you get any more light on the subject i'd be interested to know more and replace my guesses with some empirical results instead. Cheers!

An 800w motor is definately causing a spike on the power line.
I wouldn't think that would be the case with the transformer in fluorescent lights.

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

P.S.
As to above comments again i think issue is "emi" or actually reactive power trough mains lines, therefore cable shielding probably hasnt a point (as its not actually radiated). The dedicated power solution for the purpose would probably be reactive capacitors (kvar as in industrial use), but at the scale this is, probably "motor starting capacitor"(mains rated) or an emi filter can mitigate the issue.

To improve things for the entire household and prevent stuff like this in the future, a similar solution to reactive capacitors has to be sourced, but for the size of the household in question. I dont know what that is, or what type of power distribution box it has, but can look out for something that "improves cosPHI (aka power factor)" or adds "reactive capacitance" or something like that. Sry dunno the exact wording.

Industrially there are uC based solutions that simply monitors the need of reactive power and switches reactive capacitors on and off as needed. For small smps power supplies there are uC solely dedicated to improve the cosPHI (power factor) of the psu in question, mostly without using direct capacitance. But a solution for an entire household power factor is at a scale that i do not know, although i do not doubt there must be something for this scale range too. Although i dont know it, I would still expect it to be quite expensive.

Yup neons are big consummers of reactive power just like motors, or anything that shifts out current and volts out of phase to each other, usually anything with significant inductance. If neons are linked to classical transformers with big values, e.g. 125W transformers, you can easily draw more reactive power than an 800W motor. And about transients, at startup the neons heat up the ends to incandescence to start up the neon, which is a small short, then close, so they make 2 on and off ones, but that happens each time they blink. And sometimes they can blink a lot before starting properly.

Imo should check the neon instalation to see first if it does use transformers and how much total wattage. Could try replicate it with starting some big motors with the same phase in another room, then its more than clear its reactive power issue. If that is the case, could try use electronic neon starter solution instead, but that does not actually fix a bad cos phi power factor that might be there, and the problem has all chances to manifest again with the next reactive power consumer.

A good power factor fix would be good generally speaking, but what i forgot to say is that you may not want to actually fix that. Industrial consumers have both active and reactive counters and pay both kw and kwar prices. All chances are a household only has active kw counter, and depending on how your cosphi is now and how it ends up to be, you can also end up indeed paying 3-4-5% less but also there is a certain situation where you can end up paying that more. Should see what is exact situation before doing that.

I would try first an emi filter with big values components like ones dedicated for welding tools.

P.S. actually looking at them now the welding filters have big amps, therefore they need to have low component values starting with low inductors (high current gets them to be low actual inductance for small form factor), which probably makes them less effective for this application, in other words shouldnt be recommended, i missed out on this aspect, didnt thought that far.

[eu.mouser.com] and probably SCHURTER 5500.2027 at 1 or 2 amps have inductors of 10 and 4 mH which is quite big values, could be good. For diy, old TV / old pc monitor entrance emi filters those are sometimes even double inductors like a L-C-L-C filter, and should be quite good for this. ATX psu has a 1uF film non polarized cap 250V rating which can be added to this all. I think the Cy must necesarily to be Cy caps types, so if they fall they are made specially to fall open and not short out the phases to the case.

Edited 2 time(s). Last edit at 08/15/2012 11:28AM by NoobMan.

maybe you can change the fluorescent lights for some incandescent ones?