The only thing I'd disagree with is that the shields should be tied to the common. ?Everything I've been taught says the shields are connected to the frame and ultimately the Earth connection. ? If you build that faraday shield it's not connected to the DC common but to the frame/earth ground is it not? ?So bringing a shielded cable in or out of the faraday shielded enclosure means the shield is part of that overall barrier against electrically radiated interference.? The shield doesn't do much for magnetically coupled interference.?? Distance is the best solution for that. ? So in short, ideally the shield should never have any current flowing through it unless it's transferring electrical noise through ground/earth back to the source.? The moment you connect the shield to the DC common you invite current flow through it too. ? John Dammeyer ? ? ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of BuffaloJohn
Sent: December-18-19 1:25 PM
To: [email protected]
Subject: Re: [digitalhobbyist] EMI interference #RELS? There would not be any reason to tie the ground of the VFD to the RELS common, at least not that I can think of. The VFD ground is tied to the ground of the motor and to the chassis. All the RELS parts have separate common (sometimes know as the minus of the 5v supply, sometimes know as chassis ground - it isn't ground though, because you isolate it from the motor and VFD ground to the mains. The idea is to isolate the spikes on the heavy current components from the light current components. Ground is a confusing term to use, so we use it only for things that are actually attached to a ground, either the mains ground or to a rod actually driven into the ground (hence the term ground). Common is the term preferred for when a bunch of items have the same relative potential (voltage) which for the measuring and displaying gear, is the minus of the power supply. So, all things attached to low voltage supplies (you might have a 5v and a 3.3v and a 1.8v, etc.) all have their minus leg or common tied together and all the shields for the signals are also tied to the same common and that common is _not_ tied to the ground of the heavy current items. If you wish to make the measurement stuff talk to the VFD, optically isolated communications is the way to go, so that you can keep from crossing current from each system. I don't think you are trying to do this application from what I read. I hope I explained it better. If that does not make sense, please let me know, I might be able to make a drawing to show the concepts. ? John, you lost me on the "common" where is the common between a VFD and RELS and other components in the area? ? Noise can be a difficult thing to deal with, but there are a number of solutions that you can turn to. First, it is important to understand what causes the "spikes and blips". My background is over 40 years as a EE which includes years working with sensitive measurements in harsh environments, on stationary equipment and on mobile equipment, sometimes with the added issues with radio jammers or high power antennas or the worst case of lightning... The rules are that inductors (coils, motors, wires, etc.) can not change current instantaneously, capacitors can not change voltage instantaneously, and resistors can change both instantaneously. If an inductor gets a pulse of current, it is effectively turned into a pulse of voltage across the inductor. If a capacitor gets a pulse of voltage, it is effectively turned into a pulse of current.? All components in the system are made up of inductors, capacitors, and resistors - some are more prominent than others, depending on the component. Why do we care? Because we need to chase to spikes and blips... Start first with the high current items. Make sure you have large enough wires to connect to the mains. Make sure the chassis is grounded. Keep the big wires well away from the small signal components (all that measurement gear). Wires can couple to other wires and that makes a pulse on one wire translate to the other - just like a transformer - because wires are inductors. Generally, with motors, stranded is better than sold conductors, but use what you have and keeping the high current items away from the measurement gear will have the greatest effect. Now for the harder part - you need to make sure the common of the measurement gear (not ground) is isolated from the mains. Grounding all your measurement gear will not help unless you can make sure that the common for all the measurement gear is hard tied together to the same potential. Make the common be a big wires and short wires as you can make. Most of the signals in the measurement gear is very low current, so if you can ensure all is at the same common reference, your noise should be reduced. Should you tie this common to ground? If your power supply is isolated from ground, I wouldn't, because spikes on the mains ground (a motor starting/stopping or a VFD running) will translate to the common of the measurement gear. Next, you should use shielded cables between places and the shield must be tied to the common on both ends of the connection. Any inductive coupling will be mostly coupled to the shield, but both sides of the shield need to be tied to common. Next, you need to understand how these electronic calipers work. Simply, embedded in the caliper is a series of coils which are used to determine the position of the head on the slider. Remember - wires next to a coil will induce current from the wire to the coil. So - keep wires away from the caliper coils. Small signal shielded cables don't couple like big power cables. Remember, AC is a changing signal (Alternating Current), so it will couple. Lastly, there is a final solution to try if you still have noise. From what I see in the discussion, you probably won't need this, but just in case, we have the Faraday Cage. Basically (and you can look up Michael Faraday if you want to learn more), the Faraday Cage is a way to isolate electromagnetic field from outside the cage from what is in the inside of the cage. If you need to isolate sensitive low level signals from noise, you can place the circuits within a copper box and tie the copper planes of the box to the common. You can use other materials to shield, but you need good conductivity and good mechanical stability (if you move the box about or open/close it). ? On Wed, Dec 18, 2019 at 10:41 AM John Dammeyer <johnd@...> wrote: Back when I was developing my ELS I also designed this little module to change the Open Collector Optical sensor signal into differential.? Picture attached. I was going to offer them up for sale but it turns out they weren't needed.? ? A sufficiently low value pull up resistor to prevent spurious noise plus filtering on the single pulse per rev requiring an index pulse to be there for at least two internal clock ticks meant noise from my VFD (GS-1) if there doesn't interfere. ? So filter the heck out of the supply and motor leads.? Then add clamp on ferrites onto the encoder signals.? If they aren't differential change them or the encoder to provide differential. ? John ? ? ? Hi thanks Richard . I am running a single phase 240v? supply to inverter running a 3 phase motor in delta. My 5 volt system? ?is not grounded as I am using? plastic enclosures not sure about grounding the 5V. The PSU 48V? and PSU 5V and? 2 stepper drivers are all in one plastic enclusure On Wednesday, 18 December 2019, 17:03:22 GMT, Richard <edelec@...> wrote: I am no expert on EMC but it could certainly could help.
Just to determine the situation have you changed to a 3 phase motor but
running single phase into the VFD?
If so which VFD?
How good is your grounding on any part of your system?
Is your 5v 0v connection grounded? How are all of the ground wires
connected, Star?
What is the location of various parts of the system? Encoder, Spindle
motor, Arduino control, Stepper Drives and PSU.
Good luck Richard
On 18/12/2019 15:51, les higgins via Groups.Io wrote:
> Recently converted my lathe to VFD and I think? am getting EMI
> interference on my RELS as the z axis motor judders and misses some
> steps.? I thought it may be the encoder cable picking up and I
> grounded the outer sheath mesh but it didn't improve? matters. Should
> I filter the 240v line in to? the inverter as well. Any suggestions
> would be appreciated
>
--
Clausing 8520, Craftsman 12x36 Lathe, 4x12 mini lathe, 14" Delta drill press, 40 watt laser, Consew brushless DC motors and a non working 3D printer
|
